Brand Endorsers with Role Model Function: Social Media Influencers’ Self-Perception and Advertising Literacy

Sogenannte Social Media Influencer*innen (SMI) ubernehmen als Werbebotschafter, mediale Entscheidungstrager, Online-Meinungsfuhrer und Kommunikationspartner eine neue Rolle in der Kommunikationsumgebung ihrer Follower*innen. Basierend auf dem Persuasion Knowledge Model und der Konzeption von Werbekompetenz bestehend aus den Dimensionen Wissen, Einstellung und Handeln untersucht die Studie, wie SMI ihre mediale Rolle konstruieren, inwiefern sie sich ihrer Vorbildrolle fur die insbesondere jungen Follower*innen bewusst sind und wie sie mit ihren Followern*innen interagieren. Im Jahr 2019 wurden funfzehn halbstrukturierte leitfadengestutzte Interviews mit SMI in Deutschland gefuhrt, die in verschiedenen Themenbereichen tatig sind (z.B. Fitness, Mode, Reisen, Familie). Die Ergebnisse zeigen, dass SMI als werbekompetent gekennzeichnet werden konnen: Sie kennen ihre Funktionen (Rollenverstandnis: Wissensdimension) und sind sich der damit verbundenen Wirkung ihres medialen Auftritts auf ihre Follower*innen bewusst (Rolleninterpretation: Einstellungsdimension). Auf Basis dieses Wissens und des Bewusstseins wahlen sie konkret Inhalte fur die Selbstdarstellung aus, konstruieren ihr mediales Erscheinungsbild bewusst und entscheiden sich proaktiv, inwieweit sie mit ihren Follower*innen interagieren (Rollenkonstruktion: Handlungsdimension).


Introduction
Brands and companies have developed numerous strategies and methods to contact their potential target groups on social network sites (SNS) (e.g. Siegert 2013;Tropp 2016). These include their own SNS brand accounts and respective profile pages as well as paid advertisements, sponsored contributions, and electronic word of mouth (eWOM; Evans et al. 2017;de Veirman, Cauberghe, and Hudders 2017). In doing so, these companies have recognized the potential of individual SNS users who stand out through their high presence and activity on SNS and use them as brand endorsers for a wide range of products and services (e.g., Möller 2011;Nirschl and Steinberg 2018;Seeger and Kost 2019): So-called social media influencers (SMIs; i.e., individual influential SNS users with a large number of followers to whom they provide specific content) have become integrated as disseminators in brands' marketing strategies (e.g., Hudders et al. 2017;Schach and Lommatzsch 2018).
As a result, a new marketing sector has evolved (e.g., Jahnke 2018; Nirschl and Steinberg 2018;Seeger and Kost 2019). Advertisers use social media influencer marketing by offering incentives, such as payment or free products; they encourage SMIs to create and distribute brand-related creative and authentic content embedded in their profile (de Veirman, Cauberghe, and Hudders 2017). This marketing strategy has become increasingly popular, as it makes it easier to reach a large consumer segment in a relatively short time at a low cost compared to other marketing forms (e.g., Herrmann 2017; Phua, Jin, and Kim 2017). Additionally, SMIs also provide access to target groups, such as children and adolescents, who are hard to reach through conventional advertising (de Veirman, Hudders, and Nelson 2019;de Vries, Gensler, and Leeflang 2012). The most popular SNS for SMI marketing to such young target groups is Instagram (Casaló, Flavián, and Ibáñez-Sánchez 2020;Evans et al. 2017;Faßmann and Moss 2016), as the platform has become tremendously successful: On the one hand Instagram stands out in terms of its number of users (e.g., Bitkom Research 2018; Hessischer Rundfunk: Desk Kommunikation 2020), with a vast community of young users (e.g., Germany: Feierabend et al. 2020). On the other hand it is also the second-most important SNS for companies (Statista 2020b "lends itself very well for eWOM purposes because products and brands can be visually imaged and named in the caption of the photo" ( de Veirman, Cauberghe, and Hudders 2017, 799).
Taking a closer look at persuasive SMI communication on Instagram, there are two relevant actors: SMI themselves and their target group. Following Friestad and Wright's Persuasion Knowledge Model (PKM;1994), both the target (follower) and the agent (SMI) are equal parties of a persuasion attempt and draw on their respective persuasion knowledge to manage a persuasion episode. In terms of the perspective of (especially young) targets of SMI marketing (i.e., persuasive SMI advertising), academic research already exists. This includes insights regarding target's persuasion knowledge (e.g., Boerman, Willemsen, and Van Der Aa 2017) -that is, children's and adolescents' specific advertising literacy (e.g., Young 2003;Livingstone and Helsper 2006;Hudders et al. 2017) and their understanding of specific persuasive tactics of advertisers, such as celebrity endorsements (Rozendaal, Buijzen, and Valkenburg 2011).
While quite a bit is known about the target group, researchers have not explored SMIs' role as agents in great depth. So far, some research exists on SMIs from an external perspective, particularly from an economic and a communication studies perspective (e.g., Schach and Lommatzsch 2018;e.g., Jahnke 2018;Stubb, Nyström, and Colliander 2019;Enke and Borchers 2018b;. Several studies have empirically investigated the extent to which SMI marketing is effective in achieving branding results, such as positive brand attitudes and purchase intentions (e.g., de Veirman, Cauberghe, and Hudders 2017;de Jans, Cauberghe, and Hudders 2019;Schouten, Janssen, and Verspaget 2020). Moreover, studies have also investigated the legal frameworks in which SMIs operate as organizational stakeholders and brand endorsers (e.g., Brecht 2019;  as well as their different roles as strategic communicators . Despite these research results from the perspective of external perception, there is a lack of more in-depth insights on SMIs themselves. This omission is notable if one assumes that both parties involved in opinion-and behavior-related persuasiveness must possess a certain level of persuasion knowledge (Friestad and Wright 1994).
That said, what knowledge do SMIs bring to the persuasion episode, and how do they use it to pursue their goals? For the purposes of our study, we translate SMIs' advertising-related target and persuasion knowledge as advertising literacy, which consists of three dimensions Rozendaal, Opree, and Buijzen 2016): The first dimension covers "the ability to recognize and understand advertising messages", which is conceptual advertising literacy (Rozendaal et al. 2016, 74). We operationalize this dimension as SMIs' knowledge both of themselves as SMIs and of their role as opinion leaders as well as brand endorsers. The second dimension deals with the "critical attitude towards advertising" (Rozendaal et al. 2016 74). We refer to this as attitudinal advertising literacy and define it as SMIs' ability to critically reflect about their relevance to their followers and their awareness of their role model function. SMIs as agents have to circumvent this attitudinal perception in the context of their persuasion attempt. The last dimensions stresses the retrieval of knowledge Rozendaal, Opree, and Buijzen 2016). We refer to this dimension as advertising literacy performance and operationalize it as the ability to apply this knowledge -that is, how SMIs specifically construct their role by selecting certain content for publication and to what extent they orient their content to their target group. Up until now, most research on advertising literacy has focused on target groups but not on the agents. This study addresses this desideratum and adapts the concept of advertising literacy as persuasion knowledge of SMIs. We aim to explain advertising literacy from the perception of the SMIs and to gain insights into the SMIs' function as actors within the advertising industry. With this in mind, we conducted 15 in-depth interviews with German SMIs of different thematic sectors on Instagram, focusing on their role understanding, on their role construction, and on their perception of their role model function.

Theoretical Framework: Social Media Influencer as Agents in the Persuasion
Knowledge Model and Their Advertising Literacy Friestad and Wright's Persuasion Knowledge Model (PKM;1994) focuses on how both people's knowledge of persuasion, as well as their agent and topic knowledge, influence their interpretation, evaluation, and coping of persuasion attempts, whereby both parties of the interaction are included. By using PKM's considerations as our theoretical framework, we can understand SMIs as agents. They are the responsible actors in the persuasion attempt, and their followers are the influenceable targets.
Previous empirical research has often dealt with the recipient's side as the target group of persuasion attempts (e.g., Evans and Park 2015;Boerman, Willemsen, and Van Der Aa 2017;de Pelsmacker and Neijens 2012) because the emphasis of the PKM was "on the active role of the target [and therefore] the majority of the PKM-related research focuses on the better understanding the target" (M. Campbell and Kirmani 2008, 563). However, this study changes the focus and addresses the agents' side -in this case, the SMI, because it is also "important to understand the agent in terms of the agent's knowledge about the target, beliefs about the target's knowledge, and interactions with the target" (M. Campbell and Kirmani 2008, 563). Including the SMIs own perception of their function, means to empirically focus on their knowledge dimensions. In this context, Friestad and Wright (1994) name three knowledge structures brought into the persuasion attempt, and which cannot be thought of individually because they are linked to each other (see also M. Campbell  -Topic Knowledge: This knowledge structure includes the beliefs about the specific topic (e.g., prior knowledge about the product) of the persuasion attempt (Friestad and Wright 1994), which means that the "product or issue expertise would be part of topic knowledge, with experts displaying higher topic knowledge than novices" (M. Campbell and Kirmani 2008, 552 Friestad and Wright (1994) assign a unique role to this knowledge structure. It includes the knowledge available to an agent for persuasion attempts (e.g., advertiser's motives, strategies, and tactics) -both to recognize them, to manage them, and to cope with them. It serves as a kind of "schemer schema" (Wright 1985).
We employ the target and persuasion knowledge as our theoretical framework for the present study because SMIs as agents use these knowledge structures in "the development, selection, and use of persuasion attempts" (M. Campbell and Kirmani 2008; see Figure 1): -Target Knowledge: To comprehend the SMIs' perspective, this knowledge structure must be thought of as the framework of their persuasion knowledge, especially when the focus is on young followers. SMIs permeate adolescent users' everyday lives, which is becoming increasingly important due to adolescents' potentially greater susceptibility (and vulnerability) to being influenced (Bitkom Research 2018;Feierabend et al. 2020;Theunert and Schorb 2004;Miller and Prinstein 2019). Thus, better understanding the relevance of SMIs in the socialization of adolescents, in addition to traditional socialization agents, seems urgent (Riesmeyer, Sawatzki, and Hagleitner 2021). -Persuasion Knowledge: Transferred to the role of SMIs as agents in the persuasion attempt, we translate this schematic knowledge structure based on the dimensions of advertising literacy Rozendaal, Opree, and Buijzen 2016;see also Young 2003;M. Campbell and Kirmani 2008;Wright, Friestad, and Boush 2005). As SMIs are part of the social media advertising market, we assume their persuasion knowledge influences how they conceptualize and implement persuasive messages in their content. Persuasion knowledge and implementation thereby includes, first, the knowledge SMIs have about their role (conceptual advertising literacy), how they evaluate their role (attitudinal advertising literacy), and how they draw on this knowledge and evaluation for their own behavior (advertising literacy performance).

Followers (Target) Social Media Influencer (Agent)
The study's focus on persuasion knowledge as advertising literacy and on target knowledge (authors' own illustration based on Friestad and Wright's PKM, 1994).

Social Media Influencers' Knowledge about Their Roles
We first consider conceptual advertising literacy in this study, which includes knowing the advertiser's intentions, goals, and tactics. According to Rozendaal and colleagues (2011;see also Wright, Friestad, and Boush 2005), this dimension comprises several components, such as recognition of both advertisements and their source as well as the understanding of the advertiser's intent and tactics. In terms of SMIs, conceptual advertising literacy covers their knowledge about these very components of the persuasion attempt as well as their knowledge of themselves as a source of advertising. As agents in the advertising persuasion attempt, SMIs must know and understand their role as advertisers to use this knowledge. Taking a closer look at the concept of SMIs (specifically on what SMIs should know about themselves), we will consider two things: a) their role as opinion leaders and b) how they take advantage of this role as brand endorsers. Essentially, SMIs are people who construct their own social media identity and master their self-presentation, thus building their brand image. Thereby, they gain other users' attention and win them over as followers (e.g., Khamis, Ang, and Welling 2017;Jin and Muqaddam 2019). In turn, all that effort increases the level of awareness of those brands with which SMIs cooperate and which they endorse via specially created content (Eunjin and Heather 2020). Thus, SMIs act as co-producers and mediators of advertising messages to their followers  (Sundermann and Raabe 2019). In this way, they unite several functions originally carried out by several different actors (Enke and Borchers 2018a). The term "SMI", thus, is associated with other actor titles and definitions, such as "creative agencies, advertising media, journalistic media, testimonial givers, and opinion leaders" (Enke and Borchers 2019, 267).
For this study, we base our conceptual description of SMIs on the opinion leader concept (Geise 2017;Leißner et al. 2014) to describe what characterizes SMIs. Opinion leaders influence others' attitudes and behaviors through their interpersonal communication and behavior (Dressler and Telle 2009). Their influence on opinionforming and decision-making processes is attributable to the fact that others ascribe opinion leaders a particular expertise (e.g., in a specific topic area); SMIs also possess certain personality traits and social and communicative skills (Katz 1957;also Geise 2017). However, it is impossible to dichotomously distinguish whether a person is or is not an opinion leader. Instead, the definition focusses on the "degree to which an individual can influence informally other individuals' attitudes or overt behavior in the desired way with relative frequency" (Rogers and Shoemaker 1971, 199).
Transferred to SMIs, they can be considered as a new type of online/digital opinion leaders, who use the advantages of SNS for their purposes (e.g., Abidin 2016; Duckwitz 2019; Khamis, Ang, and Welling 2017; Schach and Lommatzsch 2018). Considering the personality traits of SMIs as opinion leaders -their charisma and their attractiveness, thus their social media persona, as well as their credibility -is as important as their degree of extroversion (Schach 2018; see also Geise 2017). As opinion leaders, SMIs are also credited with a high degree of authenticity, which they (must) ensure their followers (Raun 2018;Casaló, Flavián, and Ibáñez-Sánchez 2020). Authenticity is, thereby, a multi-layered construct that is often connected with such attributes as genuineness, truthfulness, and originality (Molleda 2010). Furthermore, SMIs possess specific thematic knowledge and expertise that can be monomorphic or polymorphic (see Merton 1968). Thus, some SMIs are dedicated to one topic, whereas others are devoted to various topics. Finally, communicative skills are central to SMIs. They must demonstrate continuity and heightened activity while communicating with their followers (Geise 2017), a skill that Trepte and Böcking (2009) consider the basis of opinion leadership. Communicative literacy means SMIs can prepare content for specific target groups and effectively communicate and disseminate it via their respective communication channels (Schach 2018). Communicativeness is fundamental to everything that constitutes SMIs, as they must "signal accessibility, availability, presence [and] connectedness" (Raun 2018, 99). Furthermore, Schenk (1995)  SMIs consciously use all these attributes and qualities as opinion leaders to fill their role as brand endorsers, since opinion leadership is an essential element in marketing (e.g., Pöyry et al. 2019;Schouten, Janssen, and Verspaget 2020). Thus, SMIs can be understood as strategic communication actors (Borchers 2019; Enke and Borchers 2019), whom companies integrate specifically into their communication activities (Zerfaß et al. 2016). In this sense, SMIs "represent a new type of independent third party endorser who shape audience attitudes through blogs, tweets, and the use of other social media" (Freberg 2011, 90).
Enke and Borchers (2019) consider the size and quality of the SMIs' relationship network (i.e., their followers) and their various functions and tasks. The former factor is relevant because the number of people reached by SMIs is of great importance and one argument for advertising to customers in this manner (Bulkow, Urban, and Schweiger 2010). Several typologies of SMIs have been developed based on this attribute (e.g., Statista 2020a; C. Campbell and Farrell 2020). Overall, there is a consensus that SMIs with comparatively few followers -that is, "nano-influencers" (between 0 and 5/10k followers, depending on the definition) -can nevertheless function as successful advertisers. "Micro-influencers", who have between 10k and 100k followers, are also mentioned. SMIs with more than 100k followers are often called "macroinfluencers" (C. Campbell and Farrell 2020). However, a more significant number of followers does not necessarily correspond with a greater degree of influence: SMIs who are relatively less popular can be more effective endorsers when, for instance, they are experts in a relatively small field/topic and are, therefore, more engaged with their audience (de Veirman, Cauberghe, and Hudders 2017; Schouten, Janssen, and Verspaget 2020). Besides, SMIs can also be differentiated by the core platform on which they operate (e.g., 'Instagrammer' or 'YouTuber'), their thematic expertise (e.g., fitness bloggers), the origin of their popularity (e.g., if they first became popular due to their presence on SNS), or their regional focus (Enke and Borchers 2018a).
Thus, if we start from opinion leadership and brand endorsement as two role concepts as a definitional attempt and understand them as a basic level of conceptual advertising literacy as part of persuasion knowledge, the question arises to what extent influencers themselves have this knowledge about their function. To what extent is their conceptual advertising literacy developed? And to what extent are they aware of these precise roles and their characteristics and their functions? Therefore, we ask: How do SMIs perceive their roles as opinion leaders and brand endorsers (RQ 1)?

Social Media Influencers' Perception of their Role Model Function
Attitudinal advertising as a component of persuasion knowledge is understood as "having a critical attitude toward advertising" (Rozendaal, Opree, and Buijzen 2016, 74). It includes skepticism toward and belief in the appropriateness of sponsored content -that is, a kind of protective firewall to persuasion attempts (Rozendaal, Opree, and Buijzen 2016;see also, Boerman et al. 2018;D'Alessio, Laghi, and Baiocco 2009;Derbaix and Pecheux 2003). Thus, we understand that attitudinal advertising literacy is about a moral attitude toward advertising: "Moral advertising literacy reflects individuals' ability to develop thoughts about the moral appropriateness of specific advertising formats and comprises the general moral evaluations individuals hold toward these formats (e.g., advergames, brand placement, or TV commercials) and toward advertising in general, including its persuasive tactics (e.g., humor or celebrity endorsements, using personal data to customize commercial messages)" (Hudders et al. 2017, 337).
With this conceptualization, attitudinal advertising literacy also focuses on reflexive skills.
Adapted for SMIs, attitudinal advertising literacy emphasizes the ability to develop a moral attitude toward their roles as opinion leaders and brand endorsers as well as their media persona based on their conceptual knowledge about themselves. SMIs with attitudinal advertising literacy should be able to critically reflect about their roles and be aware of their relevance for the target group. Within the context of these roles -that is, through their agent behavior in terms of their high level of online activity, sizeable online following, and apparent self-confidence (Jäckel 2011) -SMIs become role models, as their followers orient themselves toward them (e.g., influencing commercial decisions by product recommendations or product placement).
While attitudinal advertising literacy on the target side comprises "low-effort, attitudinal mechanisms that can be effective in reducing children's advertising susceptibility under conditions of low elaboration" (Rozendaal et al. 2011, 344), for agents, the opposite is the case. SMIs as agents have to circumvent or breakthrough this attitudinal barrier in the context of their persuasion attempt. SMI advertising on Instagram exemplifies the native advertising approach. Unlike regular advertisers, SMIs -presenting themselves as approachable, (para)social friends -include advertisements in their content without triggering the same coping mechanisms as traditional advertising. Native advertising can surreptitiously prevent consumers from recognizing advertisements and thereby circumvent coping mechanisms. In other words, SMIs' knowledge about and attitudes toward their role(s) includes or promotes the formation of an awareness of their relevance as persuasive agents vis-à-vis their target group. To the best of our knowledge, previous research has not addressed the critical view of SMIs' role perceptions. Therefore, we ask to what extent SMIs are aware of their role model function based on their perception of their follower's ideas about their content and integration of persuasive messages (RQ 2)?

Social Media Influencers' Role Construction: Media Persona and Perceived
Relationship with Followers From conceptual advertising literacy as well as attitudinal advertising literacy comes advertising literacy performance, which is the ability to use the conceptual advertising knowledge Rozendaal, Opree, and Buijzen 2016). It covers the ability to retrieve and apply advertising-related knowledge and attitudes about advertising for individual persuasion attempts (Rozendaal, Opree, and Buijzen 2016;see also, Rozendaal, Buijzen, and Valkenburg 2012;Brucks, Armstrong, and Goldberg 1988).
If SMIs know what constitutes their role and have an attitude toward this role in terms of their role model function, they can actively perform their role(s) on this basis. For them, as strategic actors in the persuasion attempt, this means how they construct their media persona and how they perceive their relationship to their audience, e.g., their followers. For this conscious role construction, they use the functions they hold as opinion-leading brand endorsers. Opinion leaders have been ascribed several functions that are also evident among SMI in the online context (Geise 2017).
A key factor for influencers (and their success on SNS) is their content and their self-presentation. To fulfil their diverse tasks in this area -concerning content creation, production, distribution, and promotion -SMIs use their communicative skills and functions. They create and distribute brand-related creative and authentic content. That comprises the information function, i.e., a SMI conveys information but also helps in interpreting that information (thus also demonstrating an orientation function). Importantly, this information can be used by SMIs to advertise products to their followers. At the same time, SMIs act as gatekeepers in a certain sense (e.g., Engelmann 2016; White 1950) by selecting topics and collaborations to identify and define themselves, which they then present to their followers. In this way, they represent a kind of relays to their followers; a function also attributed to opinion leaders (Dressler and Telle 2009). This, in turn, affords SMIs a legitimizing function, as they communicate and disseminate information to demonstrate its relevance to their community (Grewal, Mehta, and Kardes 2000). Consequently, by consciously selecting content and thus legitimizing it, SMIs actively fulfil various tasks, based on their performance skills. These include multi-layered role tasks, as they often combine different role functions and are characterized by border-crossing qualities (Enke and Borchers 2018b;: Some SMIs perform both as content producers and as content distributors while, at the same time, providing testimonials. Thus they combine four performance characteristics: content production, content distribution and multiplication, influence, and social relations (which brings us back to opinion leadership) (for an overview, see Sundermann and Raabe 2019). Finally, some SMIs have a great degree of visibility and popularity, and they achieve a high level of prominence based on their relationships with certain groups/topics (Marwick 2015;Khamis, Ang, and Welling 2017). This, in turn, plays into the role model function, which shows how cross-linked persuasion knowledge is.
In addition to the content, the active performance of SMIs on SNS includes their relationship with their network (i.e., with their followers). Because SMIs can rely on their heightened communicative skills to interact with their followers and are technically approachable through various SNS, they have enhanced and multiple opportunities to engage in dynamic relationships with their followers. These relationships can culminate in intensive parasocial interactions between SMIs and their followers (Hartmann 2017). Thereby, parasocial interaction theoretically implies that media users are under the illusion of participating in mutual interactions with media figures (Leißner et al. 2014). This type of parasocial interaction, typified by perceived intimacy and imaginary relationships that play out on interactive platforms, such as Instagram, is a striking feature unique to SMIs (Jin 2018;Jin and Muqaddam 2019). More specifically on Instagram, application features, such as stories or posts, give followers the impression that they can interact with 'their' SMI (e.g., using an active form of address or the opportunity to comment on content). Consequently, based on several parasocial interaction situations, a relationship pattern emerges, based on which a connection to the media figure -in this case, SMIs -develops (e.g., Schramm and Hartmann 2010). For SMIs, parasocial relationships with their followers are mainly created by their SNS activity's intensity and frequency. Therefore, SMIs can be understood as parasocial opinion leaders -that is, opinion leaders who engage in various parasocial relationships (Leißner et al. 2014).
What makes these parasocial relationships between SMI and their followers particularly interesting and unique is the opportunity for mutual communication via numerous features, such as Instagram's direct messaging features (text and voice) or query functions, through which real dialogues can occur. The resulting relationships can be particularly strong (Abidin 2016;García-Rapp 2017), as the effect of distance is minimized by authentic dialogue (Raun 2018). Followers may even experience a kind of friendship with SMIs, such that the latter can assume the "role of an admired friend who supports them, who also shows herself vulnerable and needs their support" (García-Rapp 2017, 17). Following the idea of parasocial interaction, however, these friendships remain essentially one-sided (Lee and Watkins 2016).

Pädagogik
Zeitschrift für Theorie und Praxis der Medienbildung Medien www.medienpaed.com > 25. 07.2021 From this, we can conclude that SMIs can partly use their persuasion knowledge to construct their media persona and their (para)social relationship by using special features based on their advertising literacy skills. Therefore, we are interested in the specific knowledge SMIs use in this regard and the extent to which they are aware of how they actively construct their media persona, how they perceive their audience, and how they construct their relationship. Thus, we ask how do SMIs construct their media persona and their relationship to their followers (RQ 3)?

Method: Qualitative Interviews with Social Media Influencers
To answer our research questions, we chose a qualitative approach. Between June and August 2019, we conducted semi-structured interviews with 15 SMIs. 1 This methodological design is well-suited for examining mindsets, opinions, knowledge, and behaviors (Flick 2021;Meyen et al. 2019;Springer et al. 2015) because these perceptions are often complex and subconscious. As a result, it is often difficult to quantify these opinions and perceptions. Instead, this approach allows us to openly capture the self-perceptions of the SMIs, without a priori focusing on a particular expression or manner. It offers more time and space to reflect openly on their self-perceptions and resulting actions. Besides, the choice of this methodological approach yielded two additional benefits. First, the interviews were based on an interview guideline, which ensured that identical topics were discussed in each interview so that the interviews could be compared. Second, the semi-structured format allowed the respondents a certain level of openness and the freedom to adapt their responses to each question (Keuneke 2017;Springer et al. 2015).

Participants
To draw wide-ranging and diverse conclusions, it was important to include a broad sample of SMIs. For this reason, the term "SMI" and, thus, the prospective study participants were defined fairly broadly. Prospective interviewees needed to manage their individual and personal public account on Instagram, which needed to be sufficiently popular so that opinions and/or advertising could be spread widely to and among their followers. Thus, prospective interviewees were not required to have a certain number of followers to be included in this study. In addition, SMIs should be representative of various subject areas (e.g., lifestyle, health, fitness, fashion, etc.) to gain insights into the most diverse forms of SMIs' self-perception as possible. We recruited all of the participants through the 'direct messages' feature of the social media network Instagram. We chose this recruitment process because the prospective interviewees should be familiar with this SNS given their regular use (Meyen et al. 2019). Additionally, we used direct messaging to ensure a more personal approach and to increase the probability of an interview agreement -accordingly, their management was not contacted. We contacted the SMIs via private accounts and were transparent about our affiliation with LMU Munich and its work. For this purpose, in order to react to the functionality of the algorithms on Instagram, we began interacting with the SMIs before their planned recruitment via 'likes' and other content reactions. We finally made contact in response to a current 'Instagram story' so that the message would stand out among the likely numerous messages in the SMIs' inboxes. The majority of the interviews were conducted by telephone, with each interview lasting between 30 and 60 minutes. Some interviewees chose not to do the telephone interview, so they were instead interviewed via direct message on Instagram.
To select the sample, we looked for German-speaking SMIs who have a publicly accessible account. In order to have the greatest possible variety of different topics and reach, we did not consciously select according to these two criteria. In total, we researched 45 SMIs with whom we made contact. We did this by following them with our official accounts, liking posts and responding to them on story shares. By doing this, we wanted to increase the chance of appearing in their personal DMs. The goal of the recruitment was not a complete mapping of SMIs' topics or types; we wanted a sample as accessible as possible to get first insights into the perception of SMIs themselves.
Of the 45 SMIs contacted via Instagram, 15 agreed to be interviewed (see Table  1). Of those interviewed, ten were female, and five were male, all aged 21 to 36 years. The sample included accounts that ranged from 7k to 184k followers: two macro-(>100k), twelve micro-(10-100k), and one nano-SMI (<10k). Besides, the respondents were active in various subject areas, such as lifestyle, health, fitness, and fashion but also research and science. Eight respondents worked full time as SMIs (i.e., they were not employed elsewhere).

Interview guideline
The interview guideline was theory driven (Meyen et al. 2019). The first guideline category served as an introduction to the interview and contained questions on private social media usage -more precisely, on the respondents' SNS choices and motivations for these choices. The following parts of the questionnaire were dedicated to the theoretical focal points of the study: -Conceptual advertising literacy: SMIs' understanding of their profession (whether they viewed themselves as a SMI and when they became a SMI). -Attitudinal advertising literacy: recognition of SMIs' responsibility toward their followers as well as their critical self-perception of their role model function. -Advertising literacy performance: professional use of social media, the construction of their own social media identity, cooperation with companies (including which companies and why), and perception of the target group (to what extent they could picture their audience, how they would communicate with them, and which specific Instagram functions were used).
We transferred all the sections into 25 main questions and follow-up questions. The interviews were recorded and after that transcribed (word by word).

Data analysis
We analyzed the interview transcripts according to a theory-based approach (Meyen et al. 2019), with the dimensions of the guidelines comprising the basis of the category system used for the evaluation. The following categories were included: -everyday life: private and professional use of social media; -conceptual advertising literacy: knowledge about their respective roles as opinion leader and brand endorser; -attitudinal advertising literacy: role model function; -advertising literacy performance: role construction, construction of the relationship to their followers.
If necessary, we inductively supplemented these categories with further characteristics. All interview transcripts were read several times, and conspicuous passages of text belonging to the categories were marked. Subsequently, we discussed the assignments of the text passages to the categories and evaluated them according to a qualitative content analysis by Mayring (2015). In order to fulfill the quality criterions of intersubjective comprehensibility, we documented the entire evaluation process.

Results
Based on the qualitative interviews, we deduced three key findings. First, the findings show that the respondents perceive themselves as SMIs and have a concrete knowledge of their roles as opinion leaders and brand endorsers (knowledge such as conceptual advertising literacy). Second, the interviewed SMIs do reflect on their role model function (attitudinal advertising literacy). Third, the SMIs actively construct their media personas as well as their relationship to their target group a (advertising literacy performance) based on this knowledge and their attitude toward themselves.

Self-Perception as Opinion Leading Brand Endorsers
RQ1 analyses the extent to which SMIs perceive their roles as opinion leaders and brand endorsers and the extent to which they view themselves as SMIs. The results were ambivalent. While some legitimized their role as SMIs and their corresponding activities as a profession, others consciously distanced themselves from the term "influencer". One male interviewee remarked that due to his main job, he did not perceive himself as an SMI: "I am a [occupational field] who also has an Instagram account. But, you know, I'm not primarily an influencer, and I don't see myself as such personally".
[id110]. Other interviewees stated that they became an SMI unintentionally, e.g.: "I'd say I somehow slipped into being an influencer). [...] Why shouldn't you earn a little bit of money with your hobby?" [id109].
Nevertheless, they seemed to reflect on their extraordinary job situation and were aware that they could not compare their occupation as SMIs to a traditional job, e.g.: "And then you get a message like, 'Oh my god, this is not a real job. I go to work eight hours a day.' And stuff like that. They don't take my job seriously. Of course, you can't compare my job with an office job or anything like that. I'm honest. But I just have something creative" [id103].
Alongside their reflections on their self-perception, the interviewees also seemed to have thoroughly thought about their role as opinion leaders, e.g.: "Because I am also asked for my opinion from time to time -and I also ask for an opinion" [id104].
This self-perception includes both the topics SMIs relate to such, as fitness, as well as questions about product placements, personal questions about their lives, or even asking for advice. Overall, there was broad consensus that it was extremely important to be aware of the possible effects of their online behavior, regardless of whether or not they considered themselves to be a SMI. They also appeared to be aware of their influence on their followers' purchase intentions and attitudes (i.e., their role as brand endorsers). This finding can be seen, for example, in the following statement: "Often I also get messages that someone has bought something I have advertised for and also thinks this product is great. Of course, I am pleased about that. You can see that the work is worth it" [id109].
Nevertheless, SMIs pursued different goals: On the one hand, some had monetary objectives, such as discount codes, e.g.: "The companies will then tell you how often your discount code has been redeemed. That means either it's worth working with you or it's not" [id103].
On the other hand, occasionally, some SMIs revealed a self-perception that exceeded the interests of the traditional advertising medium, viewing themselves as personal and moral role models. One SMI said the following: "Other influencers have something of their account, some kind of advertising cooperation or they get money for a post. I don't have any of that. I invest time, sometimes many, many hours a day, and offer information for not one cent. Quite the contrary, I'm losing money" [id110].

Critical Reflections on Role Model Function
Closely linked to the results concerning SMIs' self-perception is RQ2, which addresses the following question: To what extent do SMIs reflect on their function? SMIs were indeed aware of their potential role model function to their followers, which they formed based on their knowledge (RQ1). Furthermore, they considered how their publicly expressed opinions, self-expression, and ultimately their behavior on SNS affect their followers. This attitudinal awareness ranged from serious, e.g.: "I have a huge responsibility to my followers" [id110] to casual, e.g.: "But even if I'm not a huge role model, I still think I can be an inspiration" [id109].
The awareness of SMIs about their function as a role model concerned knowledge about their potential influence on their followers' buying behavior. They recognized that followers are repurchasing certain things from them, as one participant noted: "I'm sure some followers say: 'Wow, those pants look insanely good on her. I want those, too.'" [id105].
This economic dimension of the perceived role model function among SMIs cause them to carefully select the companies with whom they chose to cooperate. One SMI in particular, who is dedicated to a healthy, athletic lifestyle, said: "I recently received a very good request, which was also very, very well paid. But I rejected it because I would have had to advertise an alcoholic drink [...] I wouldn't promote something like that, actually" [id101].
Another highlighted his responsibility in terms of the quality of the content advertised: "Because otherwise I can push something on them, and they would buy it because of me. And if I offer them junk, they would hate me for it" [id103].
Yet taking a social perspective, the interviewed SMIs were cautious about which topics they communicated (e.g., topics related to life decisions or even strokes of fate or, generally, environmental protection). Interviewee said, for example, concerning her divorce: "I had motivated so many women and in fact, after I separated, after I told a little bit, uh, what each individual woman is actually worth, uh, a lot of them divorced, too" [id103].
Another spoke about his passion as an environmentalist and his role as a role model in this regard (independent of financial benefits through cooperation): SMIs with younger followers were especially conscious and reflective about their function as role models and, therefore, adapted their content and cooperated with companies even more consciously and actively. A male interviewee explained his responsibility like this: "For me, the worst thing would be if I recommend something and then some 16-year-old kid buys it with his hard-saved pocket money, and it's some kind of shit. I would be ashamed of myself" [id101].
They proactively addressed important issues, e.g., SMIs with a so-called lipedema: "I am a role model because I stand by my lipedema" [id105], and consciously refrained from others (e.g., drugs and alcohol). SMIs actively mentioned their young followers, if any: "I think deeply about how I communicate [about alcohol consumption] because I am also addressing people who are not yet 16 or who are not yet 18" [id111].

Construction of Media Persona and Relationship with Followers
Indeed, this awareness of their role model function played a part in constructing their role, which RQ3 dealt with. RQ3 addresses the SMIs' role performance -that is, how they construct their media persona and their relationship to their followers. All interviewees demonstrated pronounced social media awareness, based on which they carefully created their media persona. Thereby, two modes of personality construction became clear. Authenticity plays a crucial role for most of the SMIs. They sought to be as authentic as possible, which was not always due to altruistic motives but occasionally served an economic agenda, as one SMI indicated: "I also think that the more authentic a person perceives you, the more likely the person will buy a product" [id109].
Again, it was evident that these considerations strongly affected the choice of collaborative partners, particularly to ensure that the advertised products matched the SMI's media persona. SMIs actively and progressively deal with possible cooperation requests and questioned them (also regarding their role model function, RQ2), as evidenced by one respondent's statement about her chosen collaborations: Another interviewee, who calls herself a "greenfluencer" and who in particular shares content about her role as a mother and about environmentally conscious living, said: "The most important thing for me is that the product suits me. Jewelry, any fitness teas, and so on, I generally reject" [id109].
Thus, SMIs want to appear authentic in order to be able to create a closeness to their followers through this authenticity. Yet some SMIs made a conscious effort to construct a parallel world for themselves, as one said: "Social media is quite a bit of a parallel world -a snapshot" [id101].
Others found themselves maintaining a balanced media persona, one which ranged between authenticity and parallelism, reality and fantasy: " Even though all SMI perceived themselves as actively reflecting upon their role, they also recognized that their (partly young) followers had relatively low awareness of mediality. They did not think their followers understood that Instagram provides a media-specific staged reality, which is likely attributable to third-person-effect-like processes.
In addition, the data also reveal that SMIs construct a certain parasocial relationship with their followers. Concerning these relationships, SMIs appeared to be pursuing more interactive and reciprocal communications with their followers, e.g.: "There is a really good exchange with me and my followers" [id105].
For instance, SMIs actively used special features provided by Instagram to contact their followers. For example, they actively and consciously asked their followers questions (e.g., via Instagram stories and question or vote functions) and expected them to answer these questions. As SMI explained: "I ask short questions using this tool [...]. I actually always read through everything. Of course, you cannot always answer everything, but I normally read through everything" [id101]. However, the interviewees also expressed how they sought to limit or control their interactions with followers online; they also wished to avoid the impression that the interaction was one-sided (e.g., some SMIs reported that they apologized to their followers via the story function if they were unable to respond). The parasocial relationship between SMIs and followers is, therefore, not unilateral but two-sided. These parasocial relationships are consciously cultivated or conversely curbed by the SMIs for their own benefit. In doing so, SMI can decide which level of closeness they allow to their followers. Some reported that they have also developed a kind of friendship with their followers: "Sure, a few friendships have already developed. I also did sweepstakes back in Berlin, and I also went to Fashion Week with a follower" [id108].
Others rejected too much closeness: "With some, you talk more than with others or write a few lines more, but that's really it. [...] I usually don't answer messages that are too private" [id109].
Nevertheless, negative experiences play a significant role in such relationships just as they do for other personalities in the public eye, e.g.: "Actual hatred. [...] Well, I read through all this stuff, and one negative comment, I really have to admit, influences me more than a hundred positive ones" [id110].
All SMIs respondents agreed on this point and, as such, sought to maintain appropriate content and tone not only with their followers but also among their followers. However, toxic phenomena like hate speech are difficult to avoid or extinguish here and elsewhere. In this context, one interviewee made an apt comparison: " with their account. Second, this study includes the SMIs perceptive and captures their self-perceptions while previous research has more often included the audience or regulatory perception. Third, the study's theoretical basis provides a combination of the persuasion knowledge model (Friestad and Wright 1994) and the concept of advertising literacy Rozendaal, Opree, and Buijzen 2016). This theoretical approach might be discussed more deeply and concretized for further research on this topic. RQ1 focuses on conceptional advertising literacy as knowledge about the function of a SMI. It becomes clear that SMIs act as brand endorsers and are aware of their opinion-leading function. Yet they reject the term 'influencer' to some extent. Rather, they see themselves as advisors. They use their role and presence on SNS to increase the level of awareness of those brands with which they cooperate and for which they advertise via specially created content (e.g., Eunjin and Heather 2020; Lee and Watkins 2016). Thus, they produce advertising content and mediate advertising messages to their followers (e.g. Enke and Borchers 2018b).
Furthermore, the results demonstrate that SMIs are aware of the responsibility that comes from their self-perception as brand endorsers and advisors, especially when they address adolescents as a target group. Since they see themselves as role models and reflect about how their audience evaluates their content and thus also their persuasive messages (attitudinal advertising literacy, RQ2), they feel an obligation toward their followers. This finding is consistent with previous research, which emphasizes that SMIs can have an orientation function for adolescents (e.g., Bitkom Research 2018; Theunert and Schorb 2004;de Veirman, Hudders, and Nelson 2019). During adolescence, new agents become important and influential, orienting youth away from their parents and toward their peers -and, possibly, SMIs -in search of new role models (Kiener 2018;Steven 2018;Rotz and Tokarski 2020).
Finally, results show that SMIs act according to their knowledge and self-perception (performance advertising literacy, RQ 3). All interviewees demonstrated their social media awareness, based on which they carefully created their media persona. In doing so, they must appear authentic while still creating a kind of parallel world. This observation can be attributed to Goffman's concept of self-presentation, with its distinction between the front door presentation -what people should see -and the back door presentation -what should remain private (Goffman 1959). SMIs construct a coherent image of their profile, which shows a section of their everyday life. This is likely in recognition of the fact that followers prefer their SMIs to be authentic (Jin 2018). SMIs know that their followers may not always recognize this form of selfrepresentation, but that they are influenced by it nonetheless. Their interaction with their followers, which is characterized by parasocial relationships on social media (Jin 2018;Jin and Muqaddam 2019), is essential to their success as SMIs (including in monetary terms). SMIs initiate this contact, but, at the same time control, and limit it to protect their own privacy.

Pädagogik
Zeitschrift für Theorie und Praxis der Medienbildung Medien www.medienpaed.com > 25. 07.2021 As with all qualitative studies, the present research involved some limitations, especially its relatively small sample size. Consequently, the results of this study cannot be generalized. Besides, interviewing professional communicators, which SMIs also are by their own understanding, raises the problem of social desirability. It must also be noted that the respondents gave desired or especially positive answers to their respective job as SMIs.
There are two potential lines of future research that should be focused: a) It became clear that including the SMIs perspective on persuasive messages and role function is very valuable and thus should be considered to examine persuasion processes in more detail (e.g., Casaló, Flavián, and Ibáñez-Sánchez 2020;. SMIs are neither only opinion leaders nor only brand endorsers; they also assume other communicative roles. That is why it is crucial to reconceptualize the definition of online opinion leadership, as SMIs are reshaping how -especially parasocial -opinion leadership works. This leads to b) arguing that SMIs and their followers should be thought of and researched together, especially in terms of their relationship to each other (as done, for example, by Brunick et al. 2016;Hwang and Zhang 2018;Reinikainen et al. 2020;Wulf, Schneider, and Queck 2021). In fact, the SMI phenomenon is a double-sided one, which should be considered from both sides -especially in the context of the PKM, the topic of persuasive (advertising) messages, and how they are being handled (advertising literacy dimensions). Both the target group (i.e., adolescents) and the agent (i.e., SMIs) are relevant parts of persuasive communication (Friestad and Wright 1994). Future research should take this importance into account and should consider two main issues: First, it should be investigated which factors influence the behavior of both SMIs and followers, for example, by taking a close look at the importance of advertising literacy. Second, the parasocial interactions and relationships that evolve on social media, especially those with SMIs, are an increasingly common phenomenon and warrant further research.