Relationship between Advertising Disclosure, Inﬂuencer Credibility and Purchase Intention

: Understanding inﬂuencer credibility and online advertising and explaining its implications is the basis for analyzing customer purchase behavior. Novelties in digital marketing are visible in the growth of advertising through digital platforms using micro-inﬂuencers, compared to the former trend of using celebrities in creating brand awareness with the purpose to reach many customers and inﬂuence their buying decisions. The aim of this study is to examine how advertising disclosure (displayed/not displayed) affects inﬂuencer credibility, while analyzing inﬂuencer type (celebrity/micro-inﬂuencer) as a moderator variable underlying this relationship. Further, this paper investigates whether brand awareness mediates the relationship between inﬂuencer credibility and purchase intention. The questionnaire was designed and data were collected from 364 respondents using the convenience sampling method on the student population from one Croatian university. Regression analysis was performed to test the set hypothesis in SPSS using the PROCESS approach and independent sample t-test. The ﬁndings show: (1) displayed advertising status increases inﬂuencers’ credibility, and (2) this relationship is not moderated by inﬂuencer type. Moreover, (3) inﬂuencer credibility has a positive and signiﬁcant relationship with purchase intention, and (4) this relationship is mediated through brand awareness. Research results indicate the importance of advertising disclosure and inﬂuencer credibility in inﬂuencer marketing, since brand awareness created through inﬂuencers’ credibility increased by displayed advertising disclosure signiﬁcantly affects purchase intention of participants.


Introduction
Digitalization and the rise of social media usage change the way people learn, communicate and make buying decisions nowadays. The size of the global influencer marketing market is growing dynamically. In 2021, the value of this market reached a record value of USD 13.8 billion, more than twice that of the past two years (Statista 2021). For example, the five-year projection from statista.com (2021) of the amount spent on influencer marketing in the U.S.A. is that it will double till 2023; USD 2.42 billion was spent in 2019 on influencer marketing while the projection is that it will rise to USD 4.62 billion.
Companies interact with customers daily through social media influencers and the effectiveness of those interactions are the basis for creating an effective marketing strategy (Febriyantoro 2020). Research conducted among 1842 marketers showed that 37.2% of them cooperate with influencers, 68% of them consider this cooperation as effective and very effective and 88% intend to continue this cooperation (Buffer 2019). In a digital world, customers decide for themselves which information they will use or which advertisement is useful for them to make a buying decision (De Veirman and Hudders 2020). Ahmad (2018) stated that more than 90% of marketers who use advertising through social media influencers find this strategy effective.
Thus, in practice brand awareness is created through influencer endorsements and together with influencers' credibility traits, this may create a strong relationship with a follower in terms of their purchasing behavior. Further, the way of displaying the sponsorship status of a certain brand matched with the influencer's personality can affect significantly a consumer's perception of a certain product and influencer credibility. Influencers have become the key factor in creating brand awareness since the activity of influencers creates brand recognition and brand recall which leads to the intention to buy (Aaker 2009). Consequently, credibility of a source has a significant influence on followers' behavior in creating brand awareness which represents one of the basic pillars of loyalty (Jiménez-Marín et al. 2021).
Despite the significant increase in interest in influencer marketing, many of its aspects have not yet been sufficiently researched (Vrontis et al. 2021;Masuda et al. 2022;Schouten et al. 2020). These include influencer marketing success factors (Hughes et al. 2019;Lou and Yuan 2019;Renchen 2020), because, despite the explosion of social influencers, their effectiveness is still low (Hughes et al. 2019). For example, Renchen (2020) points out that the relationship between social media marketing effectiveness and brand awareness is not fully understood. This paper will explore whether disclosed or not disclosed advertising posts from influencers exert a different level of perceived credibility. This study extends past research on advertising disclosures and influencer marketing by examining whether influencer type moderates the underlying relationship among young consumers in Croatia. Past research on the effect of celebrity endorsement has been largely examined (Ladhari et al. 2020); however, there is a need to explore different effects that diverse influencer types may have in terms of consumer behaviors (Vrontis et al. 2021). Nowadays, micro-influencers have more effect on brand awareness and purchase intention than traditional celebrities, therefore we expect that influencer type will play a moderator role between advertising disclosure and influencer credibility.
Further, as only a small number of studies investigate brand awareness as a mediator (Ardiansyah and Sarwoko 2020), we will focus on the role of brand awareness as a possible mediator between influencer credibility and purchase intention due to the reason that customers are daily informed about products through social media advertising and brand awareness is created through continuous advertising (Keller 2003). As continuous advertising nowadays is happening through influencers, consequently brand awareness should be the significant mediator in this relationship.
It is expected that research results can help marketers to create marketing strategies targeting young consumers in using promotion of their brands through influencers by carefully choosing influencer types, bearing in mind their credibility traits and the way they communicate transparently with their followers.
The paper is divided into four major sections. After the introduction, we have synthesized the literature on influencer marketing, advertising disclosure, influencer credibility, brand awareness and purchase intention. The third part presents methodology and research design. In part four, research results are analyzed, and they are discussed in part five. Finally, discussion and concluding remarks with future implications are presented.

Influencer Marketing and Influencer Credibility
The rise of influencer marketing has become one of the most prominent fields in digital marketing and its adaptation in marketing campaigns seems to be an effective way to target customers (Jílková 2018). Influencer marketing represents electronic word of mouth in a digital era and most companies that engage influencers want to present their products to a large audience by decreasing costs. This is a type of marketing strategy which drives consumer brand awareness and purchase decisions through specific individuals called influencers (Scott 2015;Lou and Yuan 2019) who post their content on social media networks. It represents a type of online marketing based on the fact that followers on social networks tend to buy products or services recommended by influencers (Influencer Marketing Hub;Lee 2021). According to Ong and Ito (2019, p. 132) influencers are "opinion leaders of the new digital century" who influence followers' behaviors regarding a certain brand or a product (Freberg et al. 2011;Godey et al. 2016). The power of influencers is due to the attitudes of contemporary consumers, which "favour brands and products that incorporate social identity" (Hassan et al. 2021, p. 3) and this is nowadays encouraged by social media networks. Companies state that their financial performance is increasing thanks to the rise of influencer marketing (Ahmad 2018) and it has become a prevalent marketing strategy (Vrontis et al. 2021).
In the field of influencer marketing, five research themes have been identified by Vrontis et al. (2021) that explore the effect of social media influencers on consumer outcomes. Source characteristics and advertising disclosure fall into two categories, among others, how influencers may impact consumers' purchase behavior. Influencers may be divided into micro and macro-influencers based on the number of followers. A macro-influencer has more than 100,000 followers on one or more social media platforms, whereas a microinfluencer has up to 100,000 followers (Levin 2020). Further, they can be divided into celebrity influencers who are well known to the public via traditional media based on their professional achievements (actors, athletes, politicians) and general non-famous public influencers. The latter are "regular people" who have become "online celebrities" through sharing self-generated content on different topics and gain many followers (such as bloggers) (Lou and Yuan 2019) along the way; Schouten et al. (2020) call these influencerendorsers and classify them as micro-celebrities who interact with the follower through replies (Alperstein 2019). A micro-celebrity represents the individual who is not a wellknown supermodel or sport player (celebrity) but an individual who became "Instafamous" through self-presentation on social media platforms and can reach a mass audience by using digital pictures in creating a connection with followers (Marwick 2015).
These follower-influencer relationships lead to purchase intentions, referring to the combination of consumers' interest (his cognitive behavior) in a brand or a product and the likelihood of purchasing these items (Shah et al. 2012). When a customer has difficulties in differentiating products or services, brand awareness helps them to make a choice (Nurhayati and Hendar 2019). The actual buying behavior is based on information the follower collects (Nunes et al. 2018) about the product or service through different social media platforms. Therefore, it depends significantly on the preference for a particular product and the influence of this preference on the customer. It is closely related to consumer attitudes, and advertising can influence attitudes toward the advertised brand and therefore influence consumers' purchase intentions (Hoyer et al. 2013). Further, the theory of associative learning speaks of learning as the process of establishing connections between two phenomena. As Till et al. (2008) state, according to the results of the classical conditioning study-one common method of associative learning-using celebrities evoking positive feelings "as unconditioned stimuli in conditioning trials should lead to affectively favourable responses toward brands" (Till et al. 2008, p. 180). The endorsements of credible influencers should therefore evoke the reactions of followers that are desired from a marketing point of view. The results of the research confirmed that the level of influencer credibility positively affects the evaluation of the endorsed brand (Breves et al. 2019).
In this online interaction, source credibility represents a critical factor influencing the decision-making process, attitudes and purchasing intentions (Kim et al. 2018). When analyzing the impact of influencer marketing on social media, influencers' credibility is the key that drives or increases the value of the influencers on social media (Tsen and Cheng 2021) since it impacts followers' purchase intentions and represents the most essential antecedent to advertising effectiveness (Keller 2005). Influencer credibility may be referred to as the credibility of the influencer in the persuasion process (Lee and Kim 2020) and how well the message has been accepted by followers, and this depends on the trust and expertise of an influencer (AlFarraj et al. 2021). This theory was developed in the 1990s and describes that influencer traits consist of trustworthiness, attractiveness and expertise (Ohanian 1990). Weismueller et al. (2020) argue that the number of followers can positively affect influencers' attractiveness, trustworthiness and lead to purchase intention (Weismueller et al. 2020). When a specific influencer becomes more attractive, followers consequently share their posts further (Casaló et al. 2020) and like a snowball effect, this means they reach more followers. For brand managers this is the main goal, to reduce costs of target advertising and reach many new customers.
The dimension of influencer expertise represents influencers' ability to present products or services they advertise on social media networks. They are perceived as people who know what they are talking about so they "know the subject" (McCracken 1989). They have adequate experience and knowledge to give product reviews (Van der Waldt et al. 2009). On the other hand, trustworthiness represents followers' level of trust in the source regarding the honesty of posted content; that is, whether the follower perceives an influencers' post as their own (and honest) opinion or an opinion manipulated or influenced by someone else (Wiedmann and von Mettenheim 2021). This is especially significant to managers when companies advertise their products and instruct influencers what to say. So, the trustworthiness of an influencer message may affect followers' purchase intention regarding whether the content is biased or objective. For example, Casaló et al. (2020) found that influencers who are not paid to post product recommendations and mentions are more credible than those who do not state this relationship clearly because it triggers follower's skepticism. Attractiveness is the perception of an influencer and relates to social values (Chekima et al. 2020), representing how the follower perceives the similarity, familiarity and likeability (Ohanian 1990) of the influencer. Some authors have analyzed relatability as a credibility trait. As stated in Hassan et al. (2021, p. 5), this trait represents "accessibility, authenticity, believability, imitability, and intimacy", meaning that followers can identify with the influencer at some level when the source posts private experiences or feelings, which then builds a sympathetic relationship with followers (Hassan et al. 2021).
To reach consumers online, brand managers create strategies to impact consumers through content (Martínez-López et al. 2020). The subject of the effectiveness of advertising disclosure and influencer credibility among the student population in Croatia is scarcely documented, as is the effect of influencer type and the role of brand awareness on consumer behavior and influencer advertising actions.

Previous Research Findings
This part of the paper will explain previous research findings regarding the topic, and a summary can be found in Appendix A, Table A1 where empirical research has been presented in terms of analyzing dependent, independent, mediator, and moderator variables.

The Relationship of Advertising Disclosure and Influencer Credibility
According to the persuasion theory, followers perceive influencers' sponsored posts not as classic advertising but more as beneficial recommendations from someone they can relate to (Han et al. 2021). This relationship in the end may have a powerful effect on consumers' purchase intentions. Consumers react differently and may have positive and negative reactions in terms of straight advertising or unclear advertising posts (not being sure if the post is paid for or not). Sometimes they become sceptical (De Veirman and Hudders 2020) or critical (Van Reijmersdal et al. 2020;Han et al. 2021) when they see that the post is paid for, which may decrease or even increase influencers' credibility among consumers. Since sponsored posts and non-sponsored posts can look alike, consumers have difficulties in recognizing commercial purposes (De Veirman and Hudders 2020). Recently, there has been a rise in papers researching the influence of advertising disclosure on followers' behavior. Previous research found possible negative attitudes toward the advertising when an ad was visible compared to implicit or not disclosed (Lee and Kim 2020), and a positive attitude when sponsorship of a post was clearly stated compared to ambiguous or not disclosed, which increased ad recognition but decreased influencer credibility among young people (Vogel et al. 2020).
Therefore, to analyze if there is a difference in the perception of influencers' credibility when ads are disclosed or not we pose hypothesis H1: Hypothesis 1 (H1). Influencers' credibility will increase when influencers clearly disclose advertising status.

Moderating Effect of Influencer Type on Advertising Disclosure-Influencer Credibility Relation
The research of Vrontis et al. (2021) implies that there is a need to analyze the role of different influencer types (micro-influencer vs. macro) to see whether different consumer responses are visible. It is well known that using celebrities in advertising campaigns increases advertising effectiveness. With the rise of social media, companies increasingly started to use ordinary people, so-called micro-influencers, in promoting their brands to reach desired markets. There is little evidence in the literature about the advertising effectiveness of micro-celebrities in promoting brands in regards to the well-established evidence of celebrities advertising effectiveness. Dwidienawati et al. (2020) analyzed customers and influencers' reviews on purchase intention. The study showed that influencers' reviews positively influenced purchase intention while customers' reviews did not have a positive influence on purchase intention. Further, Fink et al. (2020) found that a celebrity endorser increased purchase intention when placing the sponsored brand image. A study from Schouten et al. (2020) found that influencer endorsement is more effective than celebrity endorsement regarding purchase intention. This category of micro-influencers possesses certain credibility characteristics that are appealing to companies as well as to their followers. Companies sponsor them because of their connectivity and relationship to their followers. Therefore, we pose the second hypothesis H2: Hypothesis 2 (H2). The positive effect of displayed advertising on influencers' credibility will be moderated by influencer type.

The Relationship between Influencer Credibility, Brand Awareness, and Purchase Intention
Previous research on the topic shows that influencers' credibility builds brand awareness (Lou and Yuan 2019) and purchasing intentions (Müller et al. 2018;Lou and Yuan 2019); however, they alone are insufficient (Pick 2021). This applies in particular to the issue of the mediation role of brand awareness discussed in this article. Thus, examining the relationship between these categories is important both from a cognitive and a practical point of view. Erdogan (1999) claims that a credible source influences consumer beliefs, opinions, attitudes and behaviors. Pick (2021) and Kosim and Pasaribu (2021) showed that influencer credibility is an important criterion for determining purchase intention (Kosim and Pasaribu 2021;Pick 2021). Previous research confirms that influencer credibility measured by influencers' expertise, trustworthiness and attractiveness has a positive and statistically significant effect on purchase intention (Weismueller et al. 2020). The same results were obtained in the studies by Kemeç and Yüksel (2021). Further, the results of the research confirmed that purchasing intentions are positively influenced by the influencers' trustworthiness and perceived expertise (Lou and Yuan 2019;Masuda et al. 2022). AlFarraj et al. (2021) found that attractiveness and expertise traits influence purchase intention through mediating role of online engagement among the aesthetic dermatology consumers in Jordan.
Research indicates that marketing strategies which include influencers need to be well adjusted according to the person's credibility traits; that is, personal attributes, characteri-zations and influencer type (Masuda et al. 2022), to give desired results such as purchase incentive.
Further, memory about an advertised brand and its recognition determines brand awareness (Febriyantoro 2020;Aaker 2014). Brand awareness can be graded from not recognizing the brand at all to the fully recalling a brand (Ilyas et al. 2020). Advertising and distribution to broad groups of customers highly influence brand awareness (Foroudi 2019), therefore its close relationship with influencer credibility should be analyzed, since influencers' credibility traits can positively create customers' desire for brand purchase (Martin-Consuegra et al. 2018). Uzunoǧlu and Kip (2014) noted that reliability of social media influencers impacts effectiveness of their communication. Renchen (2020) and Kosim and Pasaribu (2021) also showed that influencers have a positive and strong effect on the development of brand awareness. Lou and Yuan (2019) found that influencers' expertise, trustworthiness and attractiveness significantly affects brand awareness while similarity was not found to be significant. The research of Wiedmann and von Mettenheim (2021) suggest that brand managers should concentrate on influencers' trustworthiness and attractiveness, stating that the importance of expertise is insignificant since those credibility traits affect positively brand satisfaction, brand image and brand trust. The research of De Jans et al. (2020) analyzed whether the source (brand post or influencer post) of sponsored content affected advertising among Belgian adolescents. Results imply that adolescents like brand advertising more when it is through influencers' posts than through brand posts, but consequently brand awareness is higher through brand posts than through an influencer's posts.
As indicated by AlFarraj et al. (2021) in the marketing literature, influencers' credibility is associated with the effectiveness of communication, the aim of which may be brand awareness. Many prior studies show that brand awareness is a significant precedent of internet (Tan et al. 2021) and social media (Bilgin 2020; Dabbous and Barakat 2020) users' purchase intention. Significant positive relationship was found between brand awareness, brand association and perceived quality and brand loyalty and brand purchasing intention (Susilowati and Sari 2020). Research has proved that brand awareness affects the purchase intention of Thai online customers (Kosakarika 2020), Australian social media users (Arli 2017) and followers of one Indonesian influencer (Kosim and Pasaribu 2021).
To analyze the mediating effect of brand awareness, we have to test the direct effect of influencer credibility on brand awareness and purchase intention, then the direct effect of brand awareness on purchase intention, and finally to see the mediation effect (indirect effect) of brand awareness underlying this relationship, therefore we pose further hypotheses: Hypothesis 3 (H3). Influencers' credibility positively affects participants' brand awareness.

Mediation Role of Brand Awareness
The literature review showed a gap in analyzing brand awareness and its effect of mediating between influencers' credibility and purchase intention.
Previous researchers analyzed the mediating effect of brand awareness in the relationship between the attributes of mobile applications and travellers' consumer behavior, such as purchase intention and word of mouth (Boonsiritomachai and Sud-On 2020). Further, brand awareness partially mediates the relationship between corporate social responsibility and brand equity (Hafez 2018) and between content quality and brand interactivity on offline purchase intention . Further, Barreda et al. (2015) found a positive mediation effect of brand awareness between online social networks and word of mouth (WOM). Sharifi (2014) also researched the impact of consumers' cognition, affection and conation on future purchase intention through the mediation effect of brand awareness. The results indicate that brand awareness has a positive indirect effect on future purchase intention. Significant results were also found in the research of Surianto et al. (2020), where cause-related marketing (CRM) campaigns did not influence directly repurchase intentions but had a significant and positive indirect effect through brand awareness. Social networking sites' advertisement also has an effect on purchase intention only through brand awareness (Nofal et al. 2020, p. 148). On the other hand, Febriyantoro (2020) found that brand awareness does not mediate the relationship between YouTube ads and purchase intention and between social media marketing and purchase intention (Ardiansyah and Sarwoko 2020).
Therefore, an attempt has been made to address the research gap and explore the mediation effect of brand awareness in the relationship between influencer credibility and customer purchase intention through the last hypothesis: Hypothesis 6 (H6). The relationship between influencers' credibility and purchase intention is mediated by participants' brand awareness.

Methodology
In this section, we will explain the research design and methodology of the research.

Research Design
To test the research model ( Figure 1) we conducted two types of analysis. First, moderation analysis (Model 1, 5000 samples) to test the influence of advertising disclosure on influencer credibility moderated by influencer type (H1 and H2). Then, simple mediation analysis (Model 4, 5000 samples) to test the mediation effect of brand awareness between influencer credibility and purchase intention (hypothesis H3, H4, H5 and H6) with the help of Process v3.15 macro in SPSS v.26 (Hayes 2018). Mediation and moderation analysis represent a statistical method used in much business and social science research (Memon et al. 2018) in order to test a hypothesis about how an independent variable (variable X) transmits its effect on a dependent variable (variable Y) through a mediator (variable M) and a moderator (variable W) (Hayes 2018, p. 78 Figure  1). Path a (direct effect of INFCRED on BRAND) and path b (direct effect of BRAND on PURCIN) are also analyzed. According to Zhao et al. (2010), if the indirect effect ab is greater or smaller than zero and statistically significant, then mediation is present. Therefore, moderation analysis was appropriate to test if advertising disclosure and influencer type affected credibility, while simple mediation analysis was the appropriate solution to test whether brand awareness transmits its effect between influencer credibility and purchase intention in our research.

Data Collection and Measurement
For the purpose of conducting research, we chose a convenience sampling method. The sample consisted of students from Croatian University North. University North is the regional university in north-western Croatia and is the newest university in Croatia, dating from 2015, and today has approximately 4500 students. An online questionnaire was sent by mail to students of all departments in December 2021. The choice of this age group of respondents was justified by the fact that young people constitute the dominant groups of social media users. In 2019, 18% were 18-24 years old, 32% were 25-34 years old and 19% were 35-44 years old (Johnson 2022). Moreover, the use of social media by these users has become a habit (Gottfried and Shearer 2016).
We collected data from 371 respondents in Croatia. Four respondents answered that they did not use social media networks and three were missing data so they were ex- Our research model for testing the moderation consisted of two dichotomous variables; sponsorship display status (SPONS-independent variable) and influencer type (INFLTYmoderator variable) and one continuous dependent variable, influencer credibility (INF-CRED). Through this model we addressed further research questions: (a) Will different advertising disclosure (displayed vs. not displayed) increase influencers' credibility? (b) Does influencer type (celebrity vs. micro-influencer) strengthen the relationship between advertising disclosure and influencer credibility?
The mediation model consisted of one independent variable (influencer credibility-INFCRED), one dependent variable (purchase intention-PURCIN) and one mediation variable (brand awareness-BRAND). Through this model we addressed a further research question: (c) "Is there a statistically significant relationship between influencer credibility (independent variable) among the student population and purchase intention (dependent variable) and is this mediated by brand awareness (mediator)?
To confirm the mediation effect of brand awareness, in a simple mediation model there are two pathways by which an independent variable (INFCRED) can influence a dependent variable (PURCIN). Those paths are analyzed as direct and indirect effects of INFCRED on PURCIN. The direct effect analyzes the effect of the independent variable INFCRED on the dependent variable PURCIN, without the mediator variable, while the indirect effect measures the effect the mediator (BRAND) has on the relationship between the independent variable INFCRED and the dependent variable PURCIN (path c', Figure 1). Path a (direct effect of INFCRED on BRAND) and path b (direct effect of BRAND on PURCIN) are also analyzed. According to Zhao et al. (2010), if the indirect effect ab is greater or smaller than zero and statistically significant, then mediation is present.
Therefore, moderation analysis was appropriate to test if advertising disclosure and influencer type affected credibility, while simple mediation analysis was the appropriate solution to test whether brand awareness transmits its effect between influencer credibility and purchase intention in our research.

Data Collection and Measurement
For the purpose of conducting research, we chose a convenience sampling method. The sample consisted of students from Croatian University North. University North is the regional university in north-western Croatia and is the newest university in Croatia, dating from 2015, and today has approximately 4500 students. An online questionnaire was sent by mail to students of all departments in December 2021. The choice of this age group of respondents was justified by the fact that young people constitute the dominant groups of social media users. In 2019, 18% were 18-24 years old, 32% were 25-34 years old and 19% were 35-44 years old (Johnson 2022). Moreover, the use of social media by these users has become a habit (Gottfried and Shearer 2016).
We collected data from 371 respondents in Croatia. Four respondents answered that they did not use social media networks and three were missing data so they were excluded from further analysis. A total of 364 respondents were included in the final analysis.
The measurements for defined constructs in the research model were derived from the existing literature (Table 1).
Because of the research context, some measures were adjusted. We used a 5-point Likert scale for three constructs; INFCRED, BRAND and PURCIN. Variables SPONS and INFLTY were dichotomous. We measured advertising disclosure variable (SPONS) as 0 = not displayed and 1 = displayed. Further, variable influencer type (INFLTY) was measured as 1 = celebrity, 2 = micro-influencer.
At the beginning of the questionnaire, in the introduction part, we asked students to remember the moment when, thanks to the influencer's posting on social media, they became aware of a particular product (or service) brand that the influencer advertised on social media, which we will call (product/service X) and to bear in mind that recalled product or service X when answering all questions.

Purchase Intention (Kim et al. 2008; Dabbous et al. 2020)
Likert scale 1 to 5, where 1 is "I do not agree at all" and 5 is "I completely agree". PURCIN1: Do influencer posts on social media encourage you to buy and try a recommended product/service? PURCIN2: Do you actually buy products/services used and recommended by influencers? PURCIN3: Do you intend to buy a product/service recommended by an influencer in the future? PURCIN4: Do you recommend products (or services) to others, used and recommended by influencers? PURCIN5: When you choose between the products (or services) among the competition, do you base your decision on the influencer's recommendation?

Brand awareness (Yoo et al. 2000)
Likert scale 1 to 5 (1 = " I'm sure I don't remember" to 5 = "I am sure I remember". BRAND1: 1 Which of the following statements best describes your memory of the appearance of the advertised "X product" or the content of the advertised "X service"? BRAND2: I can quickly remember some characteristics of that advertised "product (service) X". BRAND3: I can quickly remember the symbol or logo of the advertised product/service X? BRAND4: I can recognize advertised product/service X, among other competing brands?

Influencer Credibility (Munnukka et al. 2016; Weismueller et al. 2020)
Likert scale 1 to 5 (1 = "I do not agree at all" to 5 = "I completely agree") INFCRED1: As for the influencer I follow on social media, I personally think that his/her posts on social media are honest. INFCRED2: As for the influencer I follow on social media, I personally think I can trust his/her social media posts. INFCRED3: As for the influencer I follow on social media, I personally think that the influencer is competent to make claims about the product/service X. INFCRED4: As for the influencer I follow on social media, I can easily identify with him/her. INFCRED5: As for the influencer I follow on social media, I think the influencer is attractive.

Influencer type (Schouten et al. 2020)
Remember the moment when, thanks to the influencer post on the social network, you became aware of a certain product (or service-product /service X) of a brand that the influencer advertised on social networks. What type of influencer was it? INFLTY1: The influencer was a celebrity. INFLTY2: The influencer was a micro-influencer.

Advertising disclosure (Lee 2021)
SPONS 1: The influencer did not show the info that the product (or service) was sponsored. SPONS 2: The influencer showed the info that the product (or service) was sponsored.

Results
The general information about the respondent's characteristics is presented in Table 2. As presented, 61.5% of respondents were females, and 38.5% were males. We also categorized students according to the period of time when they were born. As represented, most respondents, 73.6%, belonged to generations born 1997 and later, and 24.5% were born between 1981 and 1996, while only 1.9% of respondents were born from 1965 to 1980. According to the preference of which social media network they used the most, more than 50% of respondents answered Instagram (52.2%), followed by Facebook (24.5%) and then YouTube (17%).
Descriptive statistics ( Further, the validity of the questionnaires was tested. Table 3 represents a correlation matrix of analyzed continuous variables measured on a Likert scale.  All variables BRAND, INFCRED and PURCIN are statistically significant (p < 0.01) and the Pearson's correlation coefficients of all variables are above the defined critical value for the questionnaire sample (n = 364). We can also see that all variables have positive association. Therefore, the questionnaire is validated.
For testing reliability of the constructs, we used Cronbach's alpha. A Cronbach's alpha greater than 0.7 suggests that questionnaire design has relatively high internal consistency (Guilford 1965). As presented in Table 4, the construct brand awareness that consisted of 4 items is 0.771, influencer credibility construct (5 items) is 0.866 and purchase intention construct (5 items) is 0.907, representing a good internal consistency. An independent-sample t-test was run to determine if there were differences in the perceived influencer credibility (dependent variable) when an influencer displayed or not (independent variable) the advertisement in a post (H1). We then discussed the moderation analyses and effects of influencer type (H2) using the simple moderation Model 1.
The results found that influencers' credibility significantly increased when the influencer displayed the advertisement in their online posts (3.45) compared to those who did not display the advertisement, t (362) = −3.416, p = 0.001. The effect size can be determined by calculating eta square (Eta 2 = 3.29). According to Cohen's d, this presents a small effect size but is statistically significant. Therefore, we can confirm our first hypothesis that posts from those who clearly displayed their sponsorship status had increased influencer credibility. Independent sample t-tests were conducted to compare female and male participants and perceived influencer credibility. The results show no significant differences between gender in influencer credibility. The female participants did not rate influencer credibility significantly higher (M = 3.27, t = 0.096, df = 362, p = 0.924) than the male participants (M = 3.28). The moderation analysis (Table 5) showed that influencer type did not moderate the relationship between sponsorship display status and influencer credibility (p = 0.369), so H2 is rejected. After determining the relationship between advertising disclosure and influencer credibility, Model 4 was applied in SPSS to see whether increased influencer credibility influenced purchase intention through brand awareness.
The regression analysis results of the model coefficient are presented in Table 6. statistically significant and has a positive relationship (beta = 0.123, p < 0.05). Path c' represents the direct effect of influencer credibility on followers' purchase intention, that is, hypothesis H5: "Influencers' credibility positively affects participants' purchase intention". This hypothesis is also supported since the relationship is positive and statistically significant (beta = 0.526, p < 0.01). Independent sample t-tests were conducted to compare female and male participants on perceived purchase intention. The results show significant differences between gender in the perceived purchase intention. The female participants displayed a significantly higher rating for purchase intention (M = 2.70, t = −2.187, df = 362, p < 0.029) than the male participants (M = 2.48). A significant difference at 1% is higher among woman respondents than male when asking: "Do influencer posts encourage you to buy or try a recommended product/service "(p < 0.001) and significant difference at 5% for the question: "Do you really buy products that are used and recommended by influencers on social networks?" (p < 0.039). To determine whether brand awareness played a significant mediation role and to confirm hypothesis H6, we present the results of testing the direct, indirect and total effects of influencer credibility on purchase intention and 95% confidence interval with the use of a bootstrapping method (Table 7) According to Demming et al. (2017, p. 82) bootstrapping is "a non-parametric approach that bypasses the problem of questionable distributional assumptions of traditional techniques and enables an accurate test of the indirect effect" and represents a more recent approach in mediation analysis than traditional ones such as the Sobel test or the causal steps method proposed by Baron and Kenny. statistically significant and has a positive relationship (beta = 0.123, p < 0.05). Path c' represents the direct effect of influencer credibility on followers' purchase intention, that is, hypothesis H5: "Influencers' credibility positively affects participants' purchase intention". This hypothesis is also supported since the relationship is positive and statistically significant (beta = 0.526, p < 0.01). Independent sample t-tests were conducted to compare female and male participants on perceived purchase intention. The results show significant differences between gender in the perceived purchase intention. The female participants displayed a significantly higher rating for purchase intention (M = 2.70, t = −2.187, df = 362, p < 0.029) than the male participants (M = 2.48). A significant difference at 1% is higher among woman respondents than male when asking: "Do influencer posts encourage you to buy or try a recommended product/service "(p < 0.001) and significant difference at 5% for the question: "Do you really buy products that are used and recommended by influencers on social networks?" (p < 0.039). To determine whether brand awareness played a significant mediation role and to confirm hypothesis H6, we present the results of testing the direct, indirect and total effects of influencer credibility on purchase intention and 95% confidence interval with the use of a bootstrapping method (Table 7) According to Demming et al. (2017, p. 82) bootstrapping is "a non-parametric approach that bypasses the problem of questionable distributional assumptions of traditional techniques and enables an accurate test of the indirect effect" and represents a more recent approach in mediation analysis than traditional ones such as the Sobel test or the causal steps method proposed by Baron and Kenny. J. Risk Financial Manag. 2022, 15, x FOR PEER REVIEW 12 of 22 statistically significant and has a positive relationship (beta = 0.123, p < 0.05). Path c' represents the direct effect of influencer credibility on followers' purchase intention, that is, hypothesis H5: "Influencers' credibility positively affects participants' purchase intention". This hypothesis is also supported since the relationship is positive and statistically significant (beta = 0.526, p < 0.01). Independent sample t-tests were conducted to compare female and male participants on perceived purchase intention. The results show significant differences between gender in the perceived purchase intention. The female participants displayed a significantly higher rating for purchase intention (M = 2.70, t = −2.187, df = 362, p < 0.029) than the male participants (M = 2.48). A significant difference at 1% is higher among woman respondents than male when asking: "Do influencer posts encourage you to buy or try a recommended product/service "(p < 0.001) and significant difference at 5% for the question: "Do you really buy products that are used and recommended by influencers on social networks?" (p < 0.039). To determine whether brand awareness played a significant mediation role and to confirm hypothesis H6, we present the results of testing the direct, indirect and total effects of influencer credibility on purchase intention and 95% confidence interval with the use of a bootstrapping method (Table 7) According to Demming et al. (2017, p. 82) bootstrapping is "a non-parametric approach that bypasses the problem of questionable distributional assumptions of traditional techniques and enables an accurate test of the indirect effect" and represents a more recent approach in mediation analysis than traditional ones such as the Sobel test or the causal steps method proposed by Baron and Kenny. The results show that influencer credibility has a significant and positive impact on brand awareness, representing path a in the mediation model. Hypothesis H3 "Influencers' credibility positively affects participants brand awareness" is supported since the effect of influencer credibility on brand awareness is positive (beta = 0.435) and is statistically significant (p < 0.01). Further, path b represents hypotheses H4: "Consumers' brand awareness positively influences participants' purchase intention", which is also statistically significant and has a positive relationship (beta = 0.123, p < 0.05). Path c' represents the direct effect of influencer credibility on followers' purchase intention, that is, hypothesis H5: "Influencers' credibility positively affects participants' purchase intention". This hypothesis is also supported since the relationship is positive and statistically significant (beta = 0.526, p < 0.01).
Independent sample t-tests were conducted to compare female and male participants on perceived purchase intention. The results show significant differences between gender in the perceived purchase intention. The female participants displayed a significantly higher rating for purchase intention (M = 2.70, t = −2.187, df = 362, p < 0.029) than the male participants (M = 2.48). A significant difference at 1% is higher among woman respondents than male when asking: "Do influencer posts encourage you to buy or try a recommended product/service" (p < 0.001) and significant difference at 5% for the question: "Do you really buy products that are used and recommended by influencers on social networks?" (p < 0.039).
To determine whether brand awareness played a significant mediation role and to confirm hypothesis H6, we present the results of testing the direct, indirect and total effects of influencer credibility on purchase intention and 95% confidence interval with the use of a bootstrapping method (Table 7)  According to Demming et al. (2017, p. 82) bootstrapping is "a non-parametric approach that bypasses the problem of questionable distributional assumptions of traditional techniques and enables an accurate test of the indirect effect" and represents a more recent approach in mediation analysis than traditional ones such as the Sobel test or the causal steps method proposed by Baron and Kenny.
The results in Table 7 show a positive direct effect of influencer credibility on purchase intention (effect = 0.526 with a t value 10.812 and p < 0.001). The results in the table show that the indirect effect of influencer credibility on purchase intention through presumed influence of brand awareness is 0.053 with a 95% bootstrap confidence interval of 0.01 to 0.10 at p < 0.05. Since the calculated confidence interval is entirely above zero, according to Hayes (2018, p. 101) this supports the conclusion that the indirect effect is positive. Therefore, hypothesis H6 is supported, meaning that the variable brand awareness mediates the relationship between influencer credibility and purchase intention at 5% significance level. This means that participants who differ by one unit in their influencer credibility opinions are estimated to differ by 0.053 units in their intention to buy. The total effect in the table represents the sum of the entire model where total effect was 0.579, with a t value of 12.924 and p < 0.001 representing a statistically significant effect (p < 0.05).

Discussion
Past research has begun investigating the influence of advertising disclosure in the area of social media influencers (Evans et al. 2017). There is a lack of empirical evidence on the student population in Croatia concerning the effectiveness of advertising disclosure in improving consumer understanding, therefore this study explored how the displayed and not displayed advertising disclosure affected perceived influencer credibility posited by source credibility theory. Furthermore, we also examined the subsequent impact of influencer credibility on purchase intention, and the mediation role of brand awareness underlying this relationship.
Our research results partially supported the assumed hypotheses. Hypothesis H1 was confirmed and showed that participants rated influencer credibility slightly higher when they showed that an advertised post was paid than those who did not display advertising status. Therefore, advertising display status affects participants' perception about influencer credibility. Hypothesis H2, regarding moderator effect of influencer type, showed that participants did not find influencer type significant in the relationship between advertising and credibility. The third hypothesis (H3) confirmed that influencer credibility positively impacted participants' brand awareness. The fourth hypothesis (H4) confirmed the direct and positive effect of brand awareness on purchase intention. The fifth hypothesis (H5), which posed that higher influencer credibility would lead to higher purchase intention, was found significant. Finally, we posed the sixth hypothesis (H6) that brand awareness played a mediator role in the relationship between influencer credibility and purchase intention, which was found to be significant and positive. The results of the research confirmed five assumed hypotheses; the exception was the H2 hypothesis, where influencer type did not play a significant moderator role.
Our study found that clear communication regarding advertising disclosure increases credibility of a source which is in line with similar results obtained in the research of Weismueller et al. (2020). The results may indicate that followers perceive influencers as more credible when they transparently display advertising status in their posts and consequently their credibility increases compared to those that do not display advertisements, so a follower is not sure whether the post is sponsored or not, which may be explained by possible skepticism of a follower when they are not sure about the credibility of an influencer's post.
In the explored relationship, influencer type did not have a significant moderator role. However, influencer type did not strengthen the relationship between advertising disclosure and influencer credibility, contrary to previous research that found influencer endorsements are more effective than celebrity endorsements (Weismueller et al. 2020).
Further, our results confirmed that influencers' credibility positively influences purchase intention. This is in line with the results of surveys conducted among social media users from Germany (Pick 2021) and Indonesia (Kosim and Pasaribu 2021). Similar results were confirmed when studying Malaysian millennials (Hassan et al. 2021) and among German millennials (Weismueller et al. 2020) and Kemeç and Yüksel (2021) among respondents from Turkey. The positive impact of certain credibility dimensions on purchasing intentions was also demonstrated by Lou and Yuan (2019) and Masuda et al. (2022) et al. 2021). It should be noted, however, that in our research we presented the aggregated impact of the credibility dimensions, i.e., trustworthiness, expertise and attractiveness. Some authors analyzed every credibility trait separately; also, there was no significant difference between male and female respondents regarding perceived influencer credibility.
The findings of the research indicate that young consumers in Croatia find influencer credibility a significant and positive factor in affecting their intention to buy products or services, and that this intention is created through brand awareness may be explained by the use of preferred media network. For example, more than half of respondents (52%) preferred Instagram and 24% preferred Facebook. The main difference among platforms was that Instagram was used for posting visual content and Facebook for sharing different types of content. Findings of Belanche et al. (2019, p. 69) explain that there is a difference in advertising effectiveness among millennials and non-millennials, indicating that Instagram stories "enhance consumer attitude toward ads and increase perceived intrusiveness, compared to [the] Facebook wall". Lee and Kim (2020) also found that highly credible brands advertised on Instagram have a positive impact on purchase intention, message credibility, attitude toward a brand and e WOM intention, which is also confirmed in our study, where displayed sponsorship advertisement increased influencer credibility and this indirectly created brand awareness among followers. Nowadays young customers are frequently exposed to advertisements through online media networks and frequent advertising creates consumers awareness of the brand, which they can recall (Keller 2003).
In our research, mediator brand awareness intervened positively and significantly between the influencer credibility and purchase intention. The findings showed that brand awareness is a significant precedent of customers' purchase intention. The positive impact of influencer credibility on customer brand awareness is consistent with the results of research by other authors. Similar results were obtained in studies conducted among internet users of various nationalities and ages: Chinese internet users and influencer followers (Tan et al. 2021), young users of social media in Turkey (Bilgin 2020), millennials in Lebanon , students in Indonesia (Susilowati and Sari 2020), Thai online customers (Kosakarika 2020), Australian social media users (Arli 2017) and followers of one Indonesian influencer (Kosim and Pasaribu 2021). The only results that differ from others are in the study from Febriyantoro (2020), conducted on millennials in Indonesia, which did not find association with purchase intention. A significantly higher difference in perception was also found among female respondents regarding desire to buy and try advertised products and really buying the product recommended by the influencer.
As presented in the literature review, the mediating role of brand awareness has been studied in different factors influencing purchase intention, such as via YouTube ads (Febriyantoro 2020), social networking site advertising (Nofal et al. 2020), social media marketing (Kosakarika 2020), ad informativeness and ad persuasiveness (Tan et al. 2021). (Our research is among the first to analyze the mediating role of brand awareness in the relationship between influencer credibility and purchasing intention, constituting a new contribution to science.

Conclusions
This research is among the first to analyze advertising effectiveness by directly comparing advertising disclosure and its effectiveness on influencer credibility moderated by influencer type (micro vs. celebrity). In this case, influencers who displayed advertising status were perceived as more credible than those who did not clearly display it. Through analysis of the literature, an increase of interest regarding advertising disclosure within influencer marketing becomes more important due to the increasing complexity of social media usage in business strategy.
Therefore, a first implication of our study is that influencers should clearly state in their posts that their advertising is paid or not since this affects their credibility in the eyes of consumers. We did not find a significant difference in the perception of participants regarding influencer type. Therefore, our recommendation would be, no matter whether the influencer is micro or celebrity, they should be sincere in their endorsements and clearly state the sponsorship since this positively affects credibility of a source. A practical recommendation is therefore to continue to use influencer endorsers in marketing campaigns since our second implication is that influencer credibility directly affects brand awareness and purchase intention. Therefore, when an influencers display clearly their sponsorship status, participants perceive them as more credible compared to those who do not display sponsorship status. The displayed brand advertising endorsed by influencer traits affects consumers' mind in the form of brand recognition and affects brand awareness, which then indirectly and significantly affects purchase intention.
Additionally, this is among the first studies to analyze brand awareness as a mediator between perceived influencer credibility and purchase intention. We explored influencer type as a moderator variable to address possible difference in the relationship between advertising disclosure and influencer credibility. The results of the research show a positive and statistically significant relationship, adding a new body of knowledge in the field of influencer marketing. The results of this study also replicate the previous research, contributing to a better understanding of the relationship between the credibility of an influencer, customer brand awareness, and purchase intention, thus addressing the gap stated in previous research (Sokolova and Kefi 2020;AlFarraj et al. 2021;Pick 2021;Vrontis et al. 2021;Masuda et al. 2022) and adds to new knowledge in this dynamically developing area of marketing activity.
Our research results are in line with previous research conducted in the field of influencer marketing. The results confirm that influencer credibility has a positive and significant influence on brand awareness. Further, brand awareness has a positive and significant impact on purchase intention, and influencer credibility affects purchase intentions of respondents.
The research results indicate that influencer credibility is something brand managers should consider when creating a brand marketing strategy since brand awareness created through influencers' traits significantly affects participants' purchase intentions. In addition, in communication policy (such as advertising disclosure), they should use various activities to increase brand awareness, because it affects the purchasing intentions of consumers in two ways: directly influencing them and indirectly increasing the effectiveness of the influencer's credibility.
The research results are particularly valuable in the context of the surveyed respondents, who were students. The rise of social media users will continue in the future and having in mind all the disruptions the pandemic brought, many people forcedly went online, especially students, who started learning online almost overnight. This continuous online activity, especially among students, leads to changing habits, from using books to e-books, from shopping in the store to shopping online, etc. This trend will continue and purchasing behaviors will change in line with this trend, therefore influencer marketing will certainly gain more attention in future research since this is a market worth billions of dollars.

Limitations and Future Research
The research was done on a convenience sample, and future studies might be conducted on a representative sample.
Further, this study was conducted during December 2021. At that time, the university offered online classes due to pandemic measures; conducting this online survey during the COVID-19 pandemic and collecting answer was not easy since the motivation of students to fill out questionnaires was found to be low, so this may have impacted the results. Therefore, we perceive motivation of students as a limiting factor.
A second limiting factor one can perceive in the study design, where we did not address a specific product or a service or conduct an experiment with stimuli, but respondents had to remember a previous situation when, thanks to the influencer's posting on social media, they became aware of a particular product (or service) brand that the influencer advertised on social media and had to bear in mind that situation when answering all questions. However, this also can be perceived as an advantage since De Veirman and Hudders (2020) state that the results sometimes may differ depending on the type of product. Addressing only one specific product in a research study may limit generalization of research results.
A final limiting factor can be seen in the research instrument of the influencer credibility construct. Since some researchers had more than five items for testing this construct, further research could be conducted bearing in mind more than five items have been used to test the influencer credibility construct. In future research it would also be useful to analyze influencer credibility traits of trustworthiness, expertise, similarity and/or attractiveness and which are significant, since some previous studies found that some credibility traits have not the same significance on purchase intention. This study analyzed the impact advertising disclosure has on perceived influencer credibility and the overall effect of influencer credibility in the context of the direct and mediated impact of brand awareness on purchase intention. In future research, it would be advisable to check the significance of individual dimensions of credibility. It would also be interesting to see how these relationships are differentiated by different product categories or age groups. Future investigation could be done to analyze whether there are differences among young customers regarding influencer credibility and the use of preferred social network (for example Instagram vs. Facebook) in creating brand awareness. As constituted by source credibility theory, influencer credibility represents the main factor of influencers' effectiveness. Brand managers will have to focus more on other factors that might have a bigger influence on the credibility of the influencer, such as psychological factors in terms of wishful identification (Hu et al. 2020)

Institutional Review Board Statement: Not applicable.
Informed Consent Statement: Informed consent was obtained from all subjects involved in the study.
Data Availability Statement: All data are shown in article and supporting data wasn't created.

Conflicts of Interest:
The authors declare no conflict of interest.   (Janssen et al. 2022)