Conveyed by Artificial Authenticity?
The Impact of Virtual Influencers
on Brand Trust
A Quantitative Study on Consumer Perceptions
Authors: Helena Märlegård & Julia Raflund
Supervisor: Erik Lundberg
Master’s Thesis in Marketing & Consumption
Spring 2024
Graduate School, School of Business, Economics and Law, University of Gothenburg, Sweden
Märlegård & Raflund (2024)
Acknowledgments
Two years together at Gothenburg School of Business, Economics, and Law is coming to an
end, and closing the chapter with this research is a great way to finish with a bang. We feel
proud of our thesis that addresses current issues about virtual influencers, AI, and
consumers' trust in brands. We would like to express our deepest gratitude to our supervisor
Erik Lundberg for his guidance, encouragement, and positive approach to our research. His
expertise and support have been helpful when conducting this research. Also, thank you
Handels!
ABSTRACT
This study examines brand trust in the context of virtual influencers. Based on a sample of
155 respondents, the research investigates the mediating effects of consumers’ perceived
authenticity and credibility of influencers in shaping brand trust. A between-subject
experiment was conducted, where influencer type was manipulated in three surveys. Findings
revealed that influencer type, whether a human, a human-like virtual influencer, or an
anime-like virtual influencer, does not directly affect consumers’ trust in brands. However,
the study underscores the importance of consumers’ perceived authenticity and credibility of
influencers in positively affecting brand trust, offering insights for marketers on strategically
leveraging influencers for desired brand trust. Consumers’ perceived authenticity of a
human-like virtual influencer was higher than for a human influencer and was fully mediating
the effect of influencer type on brand trust, contrary to an anime-like where this was not a
significant mediator. Furthermore, consumers’ perceived credibility of virtual influencers
fully mediated the level of consumers’ trust in brands, highlighting the importance of this
component. Authenticity, characterized by genuineness and alignment with intrinsic
motivations, emerges as a critical driver of consumer trust. Similarly, credibility,
encompassing trustworthiness, expertise, and attractiveness significantly influences brand
trust. Aligning brand values and product fit with influencer type is highlighted as a strategic
imperative, emphasizing the need for genuine relationships between brands and influencers to
improve consumers’ perceptions and trust in brands. The study is not free from limitations,
including challenges of manipulating the experiment, and suggests future research to explore
industry variability and a broader demographic. Overall, this research provides valuable
insights into the evolving landscape of virtual influencers, offering practical
recommendations for marketers and paving the way for further exploration into the
complexity of brand trust in the digital age.
Keywords: Brand trust, Authenticity, Credibility, Influencer marketing, Virtual influencers,
Artificial intelligence, Anthropomorphism, Uncanny valley Theory
1
Märlegård & Raflund (2024)
1. Introduction
Many scholars have argued that trust is essential for establishing successful and rich
brand-consumer relationships (Chaudhuri & Holbrook 2021; Molinillo, Ekinci & Japutra
2019; Rajavi, Kushwaha & Steenkamp 2019). According to Chaudhuri and Holbrook (2021),
brand trust positively impacts both business performance and customer loyalty, as well as
consumers’ purchase intentions and beliefs in the brand to perform as it promises. Brand trust
is a major objective in marketing and branding since it is a crucial element in relationship
building (Delgado-Ballester 2004; Molinillo, Ekinci & Japutra 2019). One way for brands to
build relationships with their consumers is through collaborations with influencers in their
marketing activities (Lee & Eastin 2021). Brands’ collaborations with influencers can
positively affect both consumer attitudes and brand performance (Lee & Eastin 2021), and
the strength of the relationship between the consumer and the influencer determines the level
of trustworthiness of the influencer (Chung & Cho 2017). Since trust is such a major part of
the interplay between brands and consumers, there is a need to further explore this topic.
Today’s technological advancements allow for innovation, such as the incorporation of
artificial intelligence (AI), virtual reality (VR), and augmented reality (AR) in marketing and
social media (Wang 2021). Wang (2021) expresses that today’s interactive marketing is being
advanced through AI, and is the fastest-growing business and comprises the interconnected
and conversable consumer, co-creating value with the different connected actors in the digital
landscape and social media. One of these technological advancements has allowed the
creation of so-called Virtual Influencers (VIs) which are artificially created humans that are
gaining popularity in today’s social media environment and act as famous influencers
(Thomas & Fowler 2021). This raises questions about their perceived authenticity (Lou et al.
2023; Sands, Campbell, Plangger, & Ferraro 2022) and credibility among consumers (Jhawar,
Kumar & Varshney 2023).
As authenticity (Ilicic & Webster 2014; Portal, Abratt & Bendixen 2019) and credibility (Lou
& Yuan 2019) are fundamental aspects of the decision-making process and trustworthiness of
brands, there is a need to explore the interplay between the increasing use of VIs, consumers’
perceptions, and the effect it has on brand trust. Most research today has examined the
perceived credibility and trust of influencers and VIs as such, or the trustworthiness of a
brand itself. Less has been studied on how the collaboration between VIs and brands affects
consumers’ trust in the brand in question. Thus, a gap in understanding the perceived
authenticity and credibility of VIs and how that affects brand trust in marketing research is
identified. As Sands et al. (2022) highlight in their study, there is a need for further research
regarding consumers’ perceived authenticity of VIs, and what qualities such as authenticity
and credibility make them successful or not (Appel, Grewal, Hadi, & Stephen 2020). It would
allow for a deeper understanding of consumer perceptions as AI is increasingly present in our
society, and in the design of VIs to keep up with this growing market.
2
Märlegård & Raflund (2024)
Rasmussen (2021) states that VIs are the future of social media marketing with
approximately three times higher engagement rates than human influencers. Additionally, VIs
are a new phenomenon that is gaining visibility and attracting new and young consumer
groups, they keep the brand at the forefront of technological advancements and offer stability
with the brand ambassador (i.e. the face of the brand). This minimizes any potential scandals
with human influencers that can threaten the brand image, putting the brand at risk
(Rasmussen 2021). Furthermore, Appel et al. (2020) argue that VIs have no limitations as to
humans who simply cannot be connected every second of the day due to the need to eat and
sleep, and believe that VIs will only continue to advance in technology and grow in presence
across social media. On the other hand, VIs are computer-based characters and can be seen as
less trustworthy (Portal, Abratt & Bendixen 2019). The collaborations between brands and
VIs can affect brand trust because brands are seen as authentic when having human-like
characters (Sands et al. 2022), which is quite paradoxical to VIs as they are crafted through
AI, and could therefore result in lower brand trust. Brands need to create innovative
marketing strategies due to the increasingly competitive marketing landscape (Jhawar, Kumar
& Varshney 2023), and several global brands have already implemented collaborations with
VIs, giving more emphasis on the dynamics of VIs and their impact on consumer behavior.
Porsche has established a collaboration with Chinese iMaker (Porsche Newsroom 2022) and
partnered up with VIs such as the Japanese @imma.gram, who is also driving the virtual
Porche Taycan (Ag 2022). In an online survey from 2022 with over 1000 American
respondents ranging from the age of 18 to 55+ years, 58% were following a VI, and 35%
stated that they had bought an item promoted by a VI (The Influencer Marketing Factory
2022). According to Garbin (2020) and Mutani (2022), the market for VIs was worth 4.6
billion dollars in 2018, having an expected compound annual growth rate of 26% by 2025.
This is a huge market expansion, highlighting the importance of VIs as a means of marketing
and business growth.
The proposed research delves into the phenomenon of VIs, and their gaining visibility and
popularity on social media. This study aims to examine brand trust in the context of VIs and
their perceived authenticity and credibility from a consumer perspective. To further deepen
the research, the influencer type, such as human influencer, human-like VI (HVI), and
anime-like VI (AVI), will be investigated to detect any potential differences in the perception
of perceived authenticity and credibility of VIs, and the effect it has on brand trust. In
summary, this research would bridge the gap between theory and practice by providing a
deeper understanding of the role VIs play in shaping brand trust. The theoretical contributions
would advance academic discussions with an insight into consumers’ perception of
authenticity and credibility regarding VIs and how this affects brand trust (Appel et al. 2020;
Lou & Yuan 201; Portal, Abratt & Bendixen 2019; Sands et al. 2022). This is an important
aspect since AI is only becoming more integral and advanced in our society. The practical
implications would guide marketers in developing more trustworthy, and therefore effective,
marketing strategies with VIs. It would also guide the design process for the creation of the
most effective type of VI, such as human-like or anime-like when marketers incorporate this
into the strategies. This would allow marketers to benefit from the growing billion-dollar
3
Märlegård & Raflund (2024)
market and the advantages of using VIs in influencer marketing. The following research
questions will guide the quantitative investigation of this research:
● How do different influencer types affect consumers’ trust in brands?
● How do the perceived authenticity and credibility of virtual influencers affect brand
trust?
2. Theoretical Framework & Literature Review
2.1 Brand Trust
With the increasing relationship marketing since the 1980s, trust has emerged as a pivotal
element within the realm of branding (Delgado-Ballester 2004). Establishing brand trust is
essential for fostering successful brand-consumer relationships and the primary objective in
marketing is to cultivate strong connections between a brand and its consumers, which builds
on trust (Molinillo, Ekinci & Japutra 2019). This will lead to customers recommending the
brand and its products and services, and by surpassing a violation, the trust can be broken and
hurt the relationship and future businesses with the brand. One factor influencing the strength
of the brand-consumer relationship is the brand personality, which Aaker (1997, p. 347)
defines as “the set of human characteristics associated with the brand”. Therefore, a brand
can be viewed as human-like or as a friend (Aaker, Fournier & Brasel 2004). Keller (1993)
believes that the brand personality and the beliefs of consumers must be congruent for the
relationship to be strong, which is also supported by Aaker (1999). In more recent research,
Leckie, Dwivedi, and Johnson (2022) found that the congruence between the self and the
brand positively impacts brand trust, aligning with previous research on brand-consumer
congruence.
Chaudhuri and Holbrook (2001, p. 83) define brand trust as “the willingness of the average
consumer to rely on the ability of the brand to perform its stated function”. Thus, consumers
trust brands to achieve what they promise to perform, which is a thoughtful cognitive process
for consumers. Moreover, high brand trust has positive outcomes on the overall business and
size of market share, due to that trust amplifies consumers’ loyalty to the brand in question
which increases their purchase frequency (Chaudhuri & Holbrook 2001). Similarly, Rajavi,
Kushwaha, and Steenkamp (2019) argue that consumers’ trust in brands affects overall brand
performance and is a determining factor of the consumer mindset of brands. Eggers,
O’Dwyer, Kraus, Vallaster, and Güldenberg (2013) believe that as part of brand trust, brand
authenticity plays a role in determining the perceived brand trust of consumers. This belief is
also supported by Portal, Abratt, and Bendixen (2019), Huang and Guo (2021), and
Schallehn, Burmann, and Riley (2014), who argue that brand authenticity positively affects
brand trust due to authentic brands carrying out what they promise and thus, staying true to
themselves and consumers. This is, as aforementioned, also supported by Chaudhuri and
4
Märlegård & Raflund (2024)
Holbrook (2001). Furthermore, successful brands are those that consumers perceive as
authentic and have human-like characteristics (Portal, Abratt & Bendixen 2019).
There are other antecedents to brand trust that could be discussed. It has been found that there
is a positive relationship between brand familiarity and brand trust ( Dawar & Lei 2009).
People are more positive towards familiar brands than unfamiliar ones since that helps them
to reduce the time for processing information about new brands, which builds better brand
trust (Ha, Pham & Le 2022). This is also supported by Akoglu and Özbek (2022), who found
previous brand experience to positively impact brand trust among consumers. Another factor
that contributes to brand trust is the emotional attachment consumers develop with the brand,
where positive emotions strengthen consumers’ trust in the brand in question (Bidmon 2017).
Bidmon (2017) suggests that a way of building these connections is for brands to create
communities and maintain positive emotions within these. As this research does not examine
consumers' direct interactions with brands per se, but instead with influencers, brand
familiarity, previous experience, or emotional attachment will not be included. Examining
previous literature on brand trust and that, as aforementioned, authenticity and credibility
strongly influence consumers’ decision-making process and level of brand trust (Ilicic &
Webster 2014; Lou & Yuan 2019; Portal et al. 2019), this research proposes that brand trust is
affected by consumers’ perceived authenticity and credibility in the context of influencer
marketing.
2.2 Influencer Marketing and Virtual Influencers
Different marketing strategies of how to best convey a message have emerged and developed
over time, from the use of traditional celebrity endorsement to influencer marketing. Lee and
Eastin (2021) found that what differentiates influencers from celebrities is the relatability to
the influencer, contrary to a celebrity. Also, Djafarova and Rushworth (2017), support this
with their finding that influencers are more influential and credible than traditional
celebrities. Influencers are famous for their activity on social media rather than for a specific
occupation and are perceived as more relatable to consumers (Meyers 2017; Silva, Farias,
Grigg & Barbosa 2020; Wang 2021). However onward, throughout this research, influencers,
and thus VIs, will be viewed and measured similarly to celebrities and celebrity endorsement,
which have been carried out in the same way by several researchers and seem to currently be
the only scale to use for this purpose (Belanche, Casaló, Flavián & Ibáñez-Sánchez 2021;
Jhawar, Kumar & Varshney 2023; Lou et al. 2023; Stein, Breves & Anders 2022).
The phenomena of VIs combine the aspects of influencer marketing and AI, and VIs are
computer-based characters that, for some, are perceived as less trustworthy, but for others as
equal to human influencers (Sands et al. 2022). There are different types of VIs, those that are
very human-like in their design and attempt to mimic humans and human behavior as closely
as feasible, and those that are less human-like and designed more as anime-like characters
(Arsenyan & Mirowska 2021). According to Molenaar (2023) and VirtualHumans (n.d.),
some of the most popular women VIs are (1) the Brazilian Lu do Magalu, @magazineluiza,
5
Märlegård & Raflund (2024)
with 6.7 million followers on Instagram, (2) the Brazilian-American Michela Sousa aka Lil
Mequela, @lilmequela, with 2.6 million followers on Instagram, (3) the Japanese Imma,
@immagram, with 393 thousand followers, (4) the English Shudu Gram, @shudu.gram, the
first digital supermodel with 240 thousand followers on Instagram, and, (5) the anime-like
German Noonoouri, @noonoouri, with 437 thousand followers on Instagram. VIs act as
“normal” human influencers on social media and are often created by large media agencies or
programmers (Stein, Breves & Anders 2022). They are precisely designed, have their own
defined personality, and create and post the same type of content as human influencers. VIs
can act as either “regular” people, or be experts in certain fields such as fashion on Instagram
or, what is known as a vTuber, for entertainment and streaming on YouTube (Stein, Breves &
Anders 2022).
To create long-term relationships with followers, influencers try to develop deep
psychological bonds (Belanche et al. 2021). Followers’ perceptions of influencers are based
on the content and promotions the influencer posts about products and brands. When there is
a mutual interest between the influencer and their followers, the strength of the relationship is
enhanced (Belanche et al. 2021). In their research, Lou et al. (2023) argue that people, in
general, follow VIs to the same extent as human influencers and that most have bought a
product recommended by an influencer due to their perceived strong relationship and
trustworthiness. On the contrary, people tend to distrust VIs due to their lack of authenticity,
and weaker relationships, and perceive them as fake since they post fabricated “human-like”
stories. Similarly, Sands et al. (2022) believe VIs not to be as relatable as human influencers
and are perceived to be socially distant, thus, providing weaker relationships. This would be
problematic due to relatable influencers usually being more successful than those who are
not. Furthermore, Lou et al. (2023) argue that one factor that contributes to the perception of
VIs to be inauthentic is the impossibility of them actually trying out a product, especially if
the product is a physical one. What could be perceived as more trustworthy is if the VI
recommends something virtual, like music, that fits with the expertise and image of the VI.
However, most people do believe that the use of VIs can help boost brands’ awareness and
add to the brand image to be, for example, trendy and up-to-date with technology (Lou et al.
2023).
2.2.1 Anthropomorphism & The Uncanny Valley Theory
Huang and Rust (2021, p. 31) conceptualize AI as “the use of computational machinery to
emulate capabilities inherent in humans, such as doing physical or mechanical tasks,
thinking, and feeling”, which take the shape either virtually or as robots. The interactions
with robots and AI in people’s daily lives are increasing, and emotions toward robots play a
big role in the willingness to interact with them (Smith, Sherrin, Fraune & Šabanović 2020).
Today, critical discussion about the social and ethical implications of AI and technical
problems possibly arising from it is detrimental (Hrastinski, Olofsson, Arkenback, Ekström &
Ericsson 2019). People are questioning AI as they previously did with robots; “will it surpass
human intelligence and trick humans?”. Nakada, Kavathatzopoulos, and Asai (2021) argue
that interactions with robots can give comfort or anxiety in different ways, but mainly people
6
Märlegård & Raflund (2024)
want robots to solve their problems without them even having to think, although humans are
the creators and therefore the ones coming up with the solutions that the robots later display.
On the other hand, Panjrath and Tiwari (2021) argue that AI-powered VIs have the ability to
reduce people's anxiety that otherwise could occur with human influencers. As seen
throughout the literature review and theoretical framework, the current research shows
contradictory results. This could be due to the fact that VIs, as a marketing tool, are a new
phenomenon and have not yet been thoroughly examined.
People tend to ascribe human attributes and characteristics, both psychological and physical,
to robots when being exposed to them in the marketplace, which is the phenomenon of
Anthropomorphism (Kim, Schmitt & Thalmann 2019). Making robots and AI technology,
like VIs, appear human-like is, according to Han and Yang (2018), crucial to driving people
to develop relationships and be willing to interact with them. They also highlight the need to
design robots with physical and psychological human traits to keep customer satisfaction and
interaction high. Dabiran, Farivar, Wang, and Grant (2024) found that physical
anthropomorphism did not affect the perceived credibility of VIs, but what affected this was
instead psychological anthropomorphism such as their moral and cognitive behavior.
However, Kim, Schmitt, and Thalmann (2019) found that when people face an
anthropomorphized robot, their perception of the robot’s warmth negatively affects their
attitudes toward the robot. That is, the more warmth the robot portrays, the more negative
attitudes people get, which sounds contradictory. This is due to the phenomenon of the
Uncanny Valley Theory, a feeling of unease for the robot or artificial entity when it gets too
close to humans in its appearance and fictional personal traits. Additionally, Kim, Schmitt,
and Thalmann (2019) state that, especially, giving a robot a human-like mind, induces
people’s feeling of uncanniness. As aforementioned, Lou et al. (2023) found that when people
know that the VI they see is, in fact, artificially created, they tend to perceive the VI as
inauthentic and distrust the VI, due to their fabricated “human-like” stories. Following
research on influencer marketing, Anthropomorphism, and The Uncanny Valley Theory, the
first hypothesis is as proposed:
H1: Influencer type has a direct effect on brand trust
2.3 Authenticity
The level of trust can be increased with perceived authenticity (Schallehn, Burmann & Riley
2014). Regardless of one’s goals, consumers seek the same three characteristics of “real”,
“true”, and “genuine”, which make up the concept of authenticity (Beverland & Farrelly
2010). Beverland and Farrelly (2010) believe that authenticity can mean different things to
different people and is a process of interpretation that varies depending on the personal goals
of the consumer. Thus, one consumer can perceive a brand or an object to be authentic, and
another to perceive the same brand or object as inauthentic. Overall, an authentic brand is
often considered to have high values and care for its consumers (Morhart, Malär,
Guèvremont, Girardin & Grohmann 2015). An authentic brand also has its identity clear,
7
Märlegård & Raflund (2024)
keeps its brand promises, and is not attributed to external forces (Schallehn, Burmann &
Riley 2014). Interactions with authentic individuals add trustworthiness in terms of what they
have demonstrated in past behavior and will make them more likely to continue behaving
predictably in the future. This reduces any risk, and thus, fosters better trust, which makes the
relationship more manageable (Schallehn, Burmann & Riley 2014).
H2: Consumers’ perceived authenticity of an influencer is positively affecting brand
trust
2.3.1 Self-Determination Theory
According to Moulard, Garrity, and Rice (2015, p. 175), influencers’, previously celebrities’,
authenticity is defined as “the perception that a celebrity behaves according to his or her true
self”, and thus, being genuine, original, and true to oneself’s feelings and beliefs (Ilicic &
Webster 2016; Moulard, Garrity & Rice 2015). Consumers perceive influencers as authentic
when the behaviors of the influencers align with the influencers' values and, indeed, stay true
to themselves (Ilicic & Webster 2016). This view of authenticity stemming from an
individual being true to oneself, is in line with the Self-Determination Theory (SDT) (Ryan &
Deci 2000). Ryan and Deci (2000) propose, according to the SDT, that authenticity is
determined by an individual behaving according to its intrinsic motivations, and thus, being
true to the self. Behaving in an intrinsically motivated manner is acting in such a way that is
truly satisfying the inner desires and beliefs. In contrast, if an individual behaves following
extrinsic motivation, this would result in actions only satisfying external desires, or acting to
gain an external reward such as compensation, which is common in influencer marketing
(Giardino 2021). Thus, intrinsically motivated behavior is perceived as authentic, whereas
extrinsically motivated behavior is deemed inauthentic (Ryan & Deci 2000). Also, Audrezet,
De Kerviler & Moulard (2020) support this view and argue that influencers behave
authentically when they collaborate with brands that align with their interests and beliefs.
This makes them behave in an intrinsically motivated manner when creating content to
promote.
H3: The perceived authenticity of an influencer is mediating the effect of influencer
type on brand trust
2.4 Credibility
Viviani and Pasi (2017) argue that credibility is an element that is perceived by individuals,
who can assess the level of credibility of sources of information they come in contact with.
The concept of credibility has been investigated over many years, and is often linked to
trustworthiness, reliability, and expertise, and can measure to what extent users think sources
are truthful, unbiased, and accurate (Viviani & Pasi 2017). The trustworthiness of an
influencer positively impacts brand credibility (Hussain 2020). Therefore, it is effective for
brands to collaborate with influencers since they are perceived as relatable and credible by
8
Märlegård & Raflund (2024)
consumers and have a positive return on investment for brands (Lee & Eastin 2021).
Consumers’ perceived credibility of an influencer positively affects their attitudes toward the
influencer in question, and in turn, the brand being promoted (Belanche et al. 2021). A
credible influencer will push consumers to react to the content shared and direct their
behavior in a desired way such as continuing to follow the influencer or doing as they
propose, as well as developing positive attitudes toward the brand being promoted. According
to Belanche et al. (2021), one crucial factor that determines the level of perceived credibility
is the fit between the influencer's image or field of expertise and the product or brand they
promoted. If these align with one another, the influencer is viewed as an expert and thus,
perceived as more credible. What could instead decrease the level of credibility is if the
influencer posts about something that does not fit as well with its image (Belanche et al.
2021). This could make consumers believe the influencer only posts for commercial
monetary benefits and therefore be biased in their recommendations, thus, decreasing their
credibility.
Previous research on this topic has also examined consumers’ perceived parasocial
relationship with VIs and the effect of disclosure vs. non-disclosure of VIs’ origin and
emotions of their content (Lim & Lee 2023). They found that when consumers are exposed to
a VI’s origin, such as that they are artificially designed, they perceive the VI as less human,
and vice versa. Furthermore, their study aimed to investigate how positive and negative
content shared by VIs affects the parasocial relationship and perceived credibility. When the
origin was disclosed, sharing positive content positively affected the relationship and
perceived credibility, and negative content had a negative effect. The outcome of the study
indicates that when VIs try to connect with consumers through negative stories, like sadness
or that something upsetting happened to them, it damages the relationship and perceived
credibility (Lim & Lee 2023). This is caused by uncanniness, such as the feeling of unease
when robots and VIs are human-like (Kim, Schmitt & Thalmann 2019). On the contrary,
positive content enhances the relationship and credibility (Lim & Lee 2023).
H4: Consumers’ perceived credibility of an influencer is positively affecting brand
trust
2.4.1 Source Credibility Theory
The Source Credibility Theory (SCT) has been widely studied for many years. It is a theory
to determine the persuasion of celebrities (McCracken 1989; Ohanian 1990), nowadays
influencers. As consumers actively take part in the “influencer world” and perceive them as
possessing symbolic meanings, influencers’ leadership status should be used also in
marketing and business (McCracken 1989). SCT consists of the perceived trustworthiness,
expertise, and attractiveness of the celebrity (Ohanian 1990), i.e. influencer. Lou and Yuan
(2019), Djafarova and Rushworth (2017), and Wiedmann and Mettenheim (2021) agree with
this viewpoint, emphasizing that expertise and attractiveness positively influence perceived
credibility. Wiedmann and Mettenheim (2021) found attractiveness to have the second largest
impact after trustworthiness, and found it also to be a determining factor in consumers’ trust
9
Märlegård & Raflund (2024)
in brands. Additionally, followers who perceive influencers as credible will also perceive
their content to be more trustworthy. Thus, the brand whose products, or overall brand, is
promoted by the influencer would face positive outcomes such as enhanced brand awareness
and increased purchase intentions (Lou & Yuan 2019). Also, Jhawar, Kumar & Varshney
(2023) posit that source credibility is a determining factor in consumers’ acceptance of VIs.
Due to previous research on credibility, the hypothesis is as follows:
H5: The perceived credibility of an influencer is mediating the effect of influencer
type on brand trust
Conceptual Research Model
We propose that the level of brand trust is mainly influenced by influencer type, with possible
mediating effects through two dimensions when collaborating with an influencer; perceived
authenticity (e.g., Busser & Shulga 2019; Eggers et al. 2013; Portal, Abratt & Bendixen
2019) and perceived credibility (e.g., Lou & Yuan 2019; Ohanian, 1990). Furthermore, due to
self-brand congruence in brand-consumer relationships and trust (Aaker 1999; Aaker,
Fournier & Brasel 2004; Leckie, Dwivedi & Johnson 2022), this research will include the
covariate fashion interest of the respondents to see any potential explanations of this on brand
trust, as this research examines brand trust in the context of fashion. Due to previous
research, the proposed research model is as follows:
Figure 1. Research Model designed by the authors
10
Märlegård & Raflund (2024)
Hypotheses Overview
Figure 2. Hypotheses Overview
3. Methodology
3.1 Approach & Objective
To explore the effect of the independent variable influencer type (X) on the dependent
variable brand trust (Y), and the potential mediating effects of perceived authenticity (M1)
and credibility (M2) of an influencer, a quantitative experimental approach was chosen
(Geuens & De Pelsmacker 2017; Hair, Black, Babin & Anderson 2019). The objective is
twofold, hence, the first part investigated how consumers perceive the authenticity and
credibility of VIs, followed by the second part that examined the impact of these perceptions
on brand trust. Due to the existing varieties among influencers (i.e. human), and in the design
of VIs (i.e. human-like and anime-like), this research will include all three types to see any
potential differences in consumers’ responses to influencers and VIs and how that affects the
level of brand trust. Consumers’ perceived authenticity and credibility of influencers were
added as mediating variables to the test. This was added to detect potential mediating effects
in the relationship between the independent and dependent variables, as a mediating variable
explains the relationship between the two, and is important in experimental research for
building new theories (Geuens & De Pelsmacker 2017; Hayes 2014).
An experimental study was conducted for several reasons, such as an objective measurement
where the collection of numerical data enables an objective analysis of consumer perceptions
and a conscientious statistical analysis of the relationships and patterns of the variables
chosen; the authenticity of the influencers, the credibility of the influencers, and brand trust
(Hair et al. 2019). Hair et al. (2019) argue that this analysis is efficient and makes it possible
to explore multiple variables and hypotheses, and is particularly popular in social science and
business to explore consumer responses to different stimuli. Since only one variable was
manipulated in each of the tests, a between-subject design was developed (Pany & Reckers
1987). Additionally, a between-subject design was the most suitable approach to not make it
obvious to respondents what the survey measured (Geuens & De Pelsmacker 2017).
11
Märlegård & Raflund (2024)
3.2 Participants & Data Collection
The participants of the study were obtained from the University of Gothenburg, targeting a
student demographic to align with the VIs and their consumer group. The students were
relatively homogenous because of their similar age, level of education, and lifestyle. Some
scholars argue that students are too different in their behavior from “regular” people and thus,
are not a good representation of the population as a whole. However, McDermott (2011)
argues that students do not differ significantly from the broader population. These differences
are believed to be too small to deem students as insufficient and therefore are approved to
participate (Geuens & De Pelsmacker 2017). Additionally, according to Geuens and De
Pelsmacker (2017), as long as there is a reasonable fit and relevance between the participants
and what is being tested, the sample is accepted.
An online survey was created since it offers the benefits of being time-efficient and
cost-effective (Duffy, Smith, Terhanian & Bremer 2005; Geuens & De Pelsmacker 2017). It
also gives the participants a chance to answer the questions at their own pace and, according
to Krumpal (2013), it also helps to generate honest answers without an interviewer being
present. The surveys were created in Qualtrix and were distributed via a QR code to
approximately 200 students during an in-class session at the University of Gothenburg, and
155 responses were recorded. The students were randomly and evenly assigned one of the
three surveys representing one influencer type each. The group presented with a human
influencer consists of 52 respondents, the group presented with an HVI consists of 53, and
lastly, the group presented with an AVI consists of 50 respondents. Randomization helps to
ensure that any differences observed between groups can be attributed to the experimental
manipulation rather than pre-existing differences between participants (Geuens & De
Pelsmacker 2017). To respect ethical considerations, the study adheres to ethical guidelines
such as gathering consent from participants, minimizing any risks associated with
participation, and ensuring the confidentiality of the respondents (Geuens & De Pelsmacker
2017).
3.3 Experiment Design
The experiment consisted of three manipulations (see Appendix 1–3), i.e. treatments (Hair et
al. 2019), where all influencers were allegedly collaborating with a big multinational fashion
and apparel brand. The brand was chosen not to be named to avoid biased connotations
(Geuens & De Pelsmacker 2017). One manipulation had a visual representation of a human
influencer, one with an HVI, and one with an AVI. All manipulations were combined with the
same written description, with the exception that the influencer was stated to be only an
influencer, a virtual one, or an anime-like virtual one. As advised by Madsen and Nielsen
(2009), the manipulations were designed as persona scenarios through storytelling, to gather
the perceptions of the participants and get a better understanding of their perceived
authenticity and credibility of the influencers. The experiment with a human influencer acts
as a counterpart to the VIs to establish a baseline of the perception of influencer authenticity
12
Märlegård & Raflund (2024)
and credibility, of influencers as such, and the effect it has on brand trust. After being shown
the manipulation, the respondents were asked to answer questions in the survey about the
VIs’ authenticity and credibility, and lastly, questions about brand trust.
Control variables were used to control what could influence the outcome and improve the
internal validity by regulating the influence of extraneous variables (Bhandari 2023). This
helped to abstain from research bias and improve relationships between the variables used
which ensured that the results were solely caused by the experimental manipulation
(Bhandari 2023). The control variables used in this research ensured that the respondent is a
frequent social media user, along with a question about their gender.
At the end of the survey, a manipulation check was included to ensure the validity of the
experiment (Kotzian, Stoeber, Hoos & Weissenberger 2020), and to check that the
manipulations were effective (Allen 2017; Geuens & De Pelsmacker 2017). It can be a
question worded directly about the manipulation, to test if the respondents perceived it in the
desired way (Allen 2017), which was the approach in this research. To test the manipulations,
three statements were included for the respondent to rate their perception on a seven-point
Likert scale ranging from 1 (strongly disagree) to 7 (strongly agree); “I perceive that this was
a human influencer”, “I perceive that this was a human-like virtual influencer”, and “I
perceive that this was an anime-like virtual influencer”. Although manipulation checks can
ensure validity, Kotzian et al. (2020) strongly emphasize not removing any participants
failing the manipulation check since this will bias the results in the way the researchers aim to
get them, leading to skewing the results of accepting or rejecting the hypotheses tested in the
research.
Figure 3. Experimental Study Design
A pre-test of the surveys was performed before the main data collection to minimize the risks
of uncertainty and potential difficulties that could occur during administrating the survey
(Hulland, Baumgartner & Smith 2018). The pre-test consisted of a small-scale study, where
the three manipulated surveys were equally distributed between 12 respondents. According to
Hulland, Baumgartner, and Smith (2018), running a pre-test is highly recommended to ensure
that the respondents understand the meaning of the questions, which will result in accurate
answers for the best result possible. The pre-test also ensured that the selected influencers on
the pictures were unknown so that no prior associations of them or undesirable or biassing
connotations were made by the respondents (Geuens & De Pelsmacker 2017). The pretest of
the manipulations was made to test how “obvious” the manipulations were and brought
13
Märlegård & Raflund (2024)
feedback on the quality of the survey design and the plausibility of the hypothesis (Geuens &
De Pelsmacker 2017).
3.4 Measurements
The measurements used in this experiment were all multi-item scales, chosen to capture the
key constructs related to perceived authenticity, credibility, and brand trust drawn from prior
research and theories. Multi-item scales are commonly used to measure constructs (Geuens &
De Pelsmacker 2017). They are believed to secure higher validity than single-item scales,
which is especially useful in empirical studies (Diamantopoulos, Sarstedt, Fuchs,
Wilczynskiv & Kaiser 2012). The items used in the experiment to test the hypothesis were all
measured on a seven-point Likert scale ranging from 1 (strongly disagree) to 7 (strongly
agree). According to Weijters, Cabooter, and Schillewaert (2010), a seven-point scale reduces
extreme responses and is also well suited when the respondents are students since they are
highly knowledgeable. Since the respondents in this research are students, a seven-point scale
was suitable.
Perceived credibility was measured with four items following that SCT is made up of
trustworthiness, expertise, and attractiveness (Ohanian 1990). The items used in the
experiment were: “This influencer is trustworthy”, “This influencer is an expert on the topic”,
”This influencer is experienced”, and “This influencer is attractive” inspired by Lim and Lee
(2023), Ohanian (1990), and Wiedmann and Mettenheim (2021). Perceived authenticity was
measured through six items; “This influencer is genuine”, “This influencer seems real to me”,
“This influencer is authentic”, “This influencer is unique in her style”, “This influencer has
distinctive characteristics”, and “This influencer has something about her that makes her
stand out”, following the belief that being authentic is to be genuine, original, and true to
oneself (Moulard, Garrity & Rice 2015). Lastly, brand trust was measured with four items; “I
trust this brand”, “I rely on this brand”, “This is an honest brand”, and “This brand is safe”,
developed by Chaudhuri and Holbrook (2021).
To have only one variable representing each of the independent, dependent, and mediating
variables, the scales were summated by merging the items measuring the same concept, into
one variable (i.e. perceived authenticity, perceived credibility, and brand trust). This
technique was used to reduce measurement errors by capturing more information when
dealing with complex concepts than what a single-item scale could have done (Hair et al.
2019). To measure the scales’ reliability, Cronbach’s alpha (𝜶) was used, which assesses
internal consistency agreed to the minimum level of 0.70 (Hair et al., 2019). All summated
scales met this criterion; for authenticity 𝜶=0.758, for credibility 𝜶=0.703, and for brand trust
𝜶=0.867 (see Appendix 4–6).
14
Märlegård & Raflund (2024)
3.5 Data Analysis
Various statistical methods in the program SPSS were used to analyze the collected data and
hypotheses. To test the hypothesis regarding the direct effect of influencer type on brand trust
(H1), the PROCESS macro was used. This is a macro used for regression analysis which
describes the statistical relations between variables, in this case, the relationship between one
dependent variable and one or more independent variables (Hayes, 2014; Xin & Xiaogang,
2009). Regression analysis can be used to examine how much a particular set of independent
variables can sufficiently explain the outcome (Xin & Xiaogang, 2009). PROCESS shows
both direct and indirect effects of an independent variable (Hayes 2014). To test for potential
differences between the influencer types, a One-way ANOVA with Tukey’s Post Hoc was
conducted. An ANOVA shows the mean score and variation between the groups compared
and thus, shows if the means are significantly different from one another (Hair et al. 2019;
Xin & Xiaogang 2009). The post hoc test was added to provide a more detailed view of the
mean differences and the magnitude of the differences between the groups (Hair et al. 2019).
The PROCESS macro was used to test if brand trust is positively affected by consumers’
perceived authenticity and credibility of an influencer (H2 & H4). Further on, the potential
mediating effects of influencer authenticity (H3) and influencer credibility (H5) were
examined, also through PROCESS. This is an additional variable that can explain the
relationship, showing the indirect effect of the independent variable on the dependent one
(Hair et al. 2019; Hayes 2014). It is suitable to include a mediating variable to test a theory
like a known outcome of an element (McKinnon 2011), such as the potential outcome of
perceived influencer authenticity and credibility. McKinnon (2011) further argues that it can
be both costly and difficult to truly measure the mediators, however, including these ensures a
more complex study, showing the differences in effects of certain elements. PROCESS
Model 4 was used as it allows for testing of mediating variables (Hayes 2014), and
bootstrapping was included for a more precise analysis of the indirect effect (Hair et al.
2019).
4. Results
4.1 Descriptive Result
A total of 186 responses were recorded, but 29 of those were incomplete. Due to these
responses missing more than 50% of the answers, they were deleted from the sample. This
resulted in a final sample size of 155 (n=155) respondents, evenly distributed across the
influencer types. Lenth (2021) states that sample sizes need to be of adequate size, such as a
moderate size that is of scientific relevance and statistically significant. The three different
treatment groups in this research consisted of 50-53 participants, which is enough to ensure
sufficient statistical power (Geuens & De Pelsmacker 2017). 54.80% of the total respondents
15
Märlegård & Raflund (2024)
were men, 40.60% were women, 3.20% were non-binary, and 1.30% preferred not to reveal
their gender. The majority of the respondents were in the age range of 19–23 years old.
As the IV is multicategorical, the group presented with the human influencer (HI group) acts
as the reference group, with two experimental groups. One presented with the HVI (HVI
group), and one presented with the AVI (AVI group). Since the HI group acts as the reference
group, the results are always in comparison to the HI group.
4.2 Manipulation Checks
Three questions aimed to validate the manipulated independent variables by asking
participants how they perceived the influencer; “I perceive that this was a human
influencer/human-like virtual influencer/anime-like virtual influencer”, rated on a 7-point
Likert-scale. The outcome indicated a significant successful validation of the manipulations
between the HI group and the AVI group (p=<0.001), but not for the HVI group (p=0.217)
(see Appendix 7). This indicates that the participants did not perceive the HVI as distinctly as
the other two categories and that the characteristics and presentation of that influencer were
not clear enough to differentiate it significantly from the other two. Since the data did not
provide strong enough evidence, the manipulation for this group was not as effective as
desired.
The F value informs the variation within and between the groups. The AVI group had the
highest score (F=46,079), the HI group had the second highest (F=10,720) and the group with
the lowest score was the HVI group (F=1,541). The F-value shows that the higher the value,
the greater the difference is between the groups, than within the group. Looking at the Sum of
Square which also shows the variation between groups, the HVI group has the lowest here as
well (8,304). The mean square within the groups (HI=2,968, HVI=2,694, AVI=2,468)
indicates homogeneity within the groups. The mean square between the groups indicates
higher differences between the HI group and the AVI group, but not the HVI group, as
aforementioned. Tukey’s Post Hoc (see Appendix 7) test complements the ANOVA results
and provides a detailed comparison between the specific groups. A combination of them both
shows a comprehensive view of the manipulation check. It confirms the successful
manipulations, as well as what should be adjusted for future research, where more support is
needed to revise the portrayal of the HVI to achieve a clearer distinction in participants’
perceptions.
16
Märlegård & Raflund (2024)
4.3 Hypothesis Testing
Figure 4. Results Overview
4.3.1 Hypothesis H1
Testing H1, with PROCESS model 4 (see Appendix 9), showed no statistically significant
direct effect of X on Y (p=0.6680 for the HVI group, and p=0,8375 for the AVI group) (see
Figure 4). Thus, the hypothesis is rejected. However, the One-way ANOVA (see Appendix
8), showed that brand trust statistically differs between the groups (F=3.417, p=0.035). To
identify individual differences between the groups, Tukey’s Post Hoc was used. The test
showed that the mean score for the group presented with the human influencer (M=3.7356,
SD=0.8216) was significantly different (p=0.032) from the group presented with the AVI
(M=3.2000, SD=1.24642), with a higher mean score. However, there was no statistically
significant difference between the group presented with the human influencer and the HVI
(M=3.5802, SD=1.08384) or between the group presented with the HVI and the AVI. Thus,
the results show a difference between two of the groups but not between all, with brand trust
scoring highest for the group presented with the human influencer compared to the AVI
group.
4.3.2 Hypothesis H2 & H4
The research model with two mediators (M1 and M2) explains approximately 35% of brand
trust (R²=0.3519) and is statistically significant (p=0.0000) (see Appendix 9 and Figure 4).
The results showed that consumers’ perceived authenticity (H2), (p=0.0010, 𝜷=0.2593) and
perceived credibility (H4) (p=0.0000, 𝜷=0.3910) of an influencer both significantly impact
brand trust positively (see Appendix 9). The results, therefore, support both hypotheses,
indicating that brand trust is positively affected by consumers’ perceived authenticity and
17
Märlegård & Raflund (2024)
credibility of an influencer, as proposed in this research. Thus, these characteristics are highly
important for influencers to portray in general regarding brand trust. The following results of
H3 and H5 show the distinct perceptions of these characteristics for the different influencer
types examined in this research.
4.3.3 Hypotheses H3 & H5
Authenticity (see Figure 4) For the HVI group, perceived authenticity is statistically
significant with a positive relationship (p=0.0068, 𝜷=0.5651) (see Appendix 9). However, for
the AVI group, perceived authenticity shows no statistical significance (p=0.5520).
Credibility (see Figure 4) For both the HVI group (p=0.0000, 𝜷=-0.9929) and the AVI group
(p=0.0000, 𝜷=-1.1850), perceived credibility is significant with a negative relationship
between influencer type and consumers’ perceived credibility.
The indirect effect of X on Y (see Figure 4) reveals a fully mediating effect through
perceived authenticity for the HVI group (95% CI [0.0152, 0.3453], not crossing 0) compared
to the HI group (see Appendix 9). However, for the AVI group perceived authenticity is
insignificant (95% CI [-0.1645, 0.0862], crossing 0), compared to the HI group. Hence, these
results partially support H3, meaning that consumers’ perceived authenticity of an HVI, but
not an AVI, is mediating the effect of influencer type on brand trust. The results also show a
significant fully mediating effect (see Figure 4) through perceived credibility for both group
HVI (95% CI [-0.6612, -0.1586], not crossing 0) and group AVI (95% CI [-0.7860, -0.1987],
not crossing 0), compared to the HI group. Thus, the results support H5, indicating that
consumers’ perceived credibility of an influencer fully mediates the effect of influencer type
on brand trust.
Figure 5. Hypotheses Results
4.3.4 Covariates
With the aim to explain brand trust even further, the covariate fashion interest was added to
the model as a direct effect on brand trust due to brand-consumer congruence positively
affecting brand-consumer relationships. With fashion interest added to the model (see Figure
4), the new explanatory power was 35.36%, with an increase of only 0.0017 decimal points
(see Appendix 10). Nor did it show a statistically significant effect on brand trust (p=0.5377).
Fashion interest did not have a statistically significant effect on brand trust in this study.
Additionally, gender was also added to the test (see Figure 4) for a potential effect on brand
trust. The results showed no statistical significance (p=0.1721), and the explanatory power
increased to 36%, indicating no major effect on brand trust.
18
Märlegård & Raflund (2024)
5. Discussion
This research explores the connection between influencer type and brand trust, offering
insights into consumer perceptions of authenticity and credibility within the context of virtual
influencers. The findings of this study show that consumers’ perceived authenticity and
credibility of influencers positively affect brand trust, which supports the proposed research
model. What did not affect brand trust directly, was what type of influencer consumers face,
indicating that simply the type of influencer does not determine consumers’ trust in brands.
Instead, brand trust is mediated by consumers’ perceived authenticity regarding an HVI but
not an AVI. Brand trust is also mediated by consumers’ perceived credibility of both VI
types. This research shows a nuanced relationship between VIs, consumer perceptions, and
brand trust.
Previous research highlights the importance of trust in creating strong brand-consumer
relationships (Delgado-Ballester 2004; Molinillo, Ekinci, and Japutra 2019) and that trust is
crucial for successful brand performance and consumer purchases (Chaudhuri & Holbrook
2001; Kushwaha & Steenkamp 2019). The results from the data set confirmed that the
perceived authenticity of an influencer positively affects brand trust, which aligns with
previous research on brand trust being positively influenced by authenticity and that this can
determine the level of brand trust (Eggers et al. 2013; Huang & Guo 2021; Portal, Abratt &
Bendixen 2019; Schallehn, Burmann & Riley 2014). Additionally, this research shows a
significant effect of consumers’ perceived credibility of influencers to affect brand trust
positively. This supports previous research that influencer credibility and trust in sources of
information affect brand trust (Belanche et al. 2021; Hussain 2020; Lou & Yuan 2019;
Viviani & Pasi 2017).
By clarifying the interplay between influencer type and consumers’ perceived authenticity
and credibility in shaping brand trust, this research contributes to influencer marketing, the
use of VIs, and brand management, offering valuable guidance for future research and
marketers when navigating this expanding and dynamic field (Appel et al. 2020; Lou & Yuan
201; Portal, Abratt & Bendixen 2019; Sands et al. 2022). The results will be further discussed
in the upcoming sections about theoretical and managerial implications.
5.1 Theoretical Implications
The findings do not support H1, that influencer type has a direct effect on brand trust,
meaning there is no direct effect. Despite this, the ANOVA with Tukey’s Post Hoc used,
showed that brand trust scored higher for the group with the human influencer than the AVI,
indicating that consumers perceive human influencers as more trustworthy than an AVI.
Furthermore, there was no significant difference between the human influencer and the HVI.
These results align with previous research emphasizing the importance of relatability and
authenticity associated with human influencers (Portal et al. 2019) and thus, reinforcing the
notion that human-like qualities play an important role in fostering brand trust.
19
Märlegård & Raflund (2024)
The Uncanny Valley Theory posits that overly human-like representations of AI and robots,
such as an HVI, can evoke feelings of discomfort (Kim, Schmitt & Thalmann 2019), but this
was not found in this research. Due to anthropomorphism, such as people attributing
human-like qualities to AI-driven technology, it could be expected that people would be more
willing to interact with an AVI than an HVI, since an HVI instead could be viewed “too
human-like” and evoke feelings of uncanniness. Thus, an AVI could be seen as more credible
(Dabiran et al. 2024; Han and Yang 2018; Kim, Schmitt & Thalmann 2019). There were no
findings of this despite the anthropomorphic nature of AVIs. The results suggest that
consumers do not exhibit any degree of anthropomorphism towards AVIs, nor did attributing
human-like qualities to the VIs foster positive outcomes among consumers, such as brand
trust. This indicates a step away from the conventional understanding of consumers’ reactions
to anthropomorphized AI-driven entities and emphasizes the intricacies of consumer
perceptions regarding virtual influencers. One potential explanation for this not influencing
consumers’ perceptions could be due to today’s digital landscape where technological
advancements are constantly increasing (Wang 2021). People are also becoming more and
more interested in VIs (Thomas & Fowler 2021), which could lead to them also being more
used to interacting with these kinds of influencers, reducing the uncanniness of the
phenomenon. This also aligns with what Pranjarath and Tiwari (2021) believe; that VIs can
reduce people's anxiety in a way that human influencers cannot.
The proposed H3, that the perceived authenticity of an influencer is mediating the effect of
influencer type on brand trust, is partially accepted. For an HVI, consumers’ perceived
authenticity of the VI fully mediates the level of brand trust. Additionally, the results even
provided a significant positive relationship between the use of an HVI and consumers’
perceived authenticity, compared to the use of a human influencer, indicating that an HVI is
perceived to be more authentic than a human influencer. What could potentially explain this
is intrinsic or extrinsic behavior, as proposed in the SDT, meaning that if an individual
behaves extrinsically, they gain an external reward which is seen as inauthentic (Giardino
2021; Ryan & Deci 2000). As humans potentially could be prone to external rewards and
monetary compensation, perhaps a VI who is computer-generated, is perceived to not have
the same external desires, and therefore, deemed as more authentic. This emphasizes the
importance of influencers behaving true to their intrinsic motivations and values. However,
the results also show that this alignment alone does not guarantee perceived authenticity.
Consumers’ perception of influencers’ authenticity consists of them being genuine, original,
true, and behaving in line with their values (Ilicic & Webster 2016; Moulard, Garrity & Rice
2015), and believe them to be authentic when they collaborate with brands suited for their
interest (Audrezet, De Kerviler & Moulard 2020). These results indicate that how unique and
well-suited the influencer is to the brand it collaborates with, is what affects consumers’
perception of influencer authenticity. Hence, the use of AI and an entity being artificially
crafted, do not seem to be a determining factor in the perception of someone being authentic
or not. This provides contradictory results to previous research by Lou et al. (2023) who
propose that when people know that a VI is artificially crafted, they believe the VI to be
inauthentic and untrustworthy. The methodology and sample of the research by Lou et al.
(2023) differ from this research. The majority of their interviewees were Chinese women who
20
Märlegård & Raflund (2024)
had previous knowledge and were following a VI on social media. This could be a reason for
the results differing from this research where previous knowledge of VIs was not a
requirement for being part of the experiments, and thus perhaps, representing a broader
picture of consumers’ perceptions.
The results of this research confirm H5, that consumers’ perceived credibility of an influencer
fully mediates the effect of influencer type on brand trust. Trustworthiness, expertise, and
attractiveness determine the source’s credibility, as proposed in the SCT (Djafarova &
Rushworth 2017; Lou & Yuan 2019; Ohanian 1990; Wiedmann & Mettenheim 2021), which
influences consumers’ acceptance of VIs (Jhawar, Kumar & Varshney 2023). The fit between
the influencer and the brand it collaborates with also affects consumers’ perceived credibility
of the influencer. If these align, the credibility increases (Belanche et al. 2021). As
aforementioned about authenticity, the experiment in this research where the influencer, being
a fashionista, was well suited to the fashion brand, seemed to positively impact the
authenticity. However, the results for credibility showed that consumers’ perceived credibility
of a VI, either being human-like or anime-like, significantly impacted the perceived
credibility negatively, compared to a human influencer. Hence, consumers do not find VIs
credible, affecting the acceptance of VIs, and thus, brand trust. This could stem from the fact
that VIs cannot actually try out a physical product, in this case, a new clothing collection. For
a VI to promote a physical product can negatively affect the trustworthiness of the VI,
contrary to if the VI would promote something virtual within its expertise (Lou et al. 2023).
This would, in turn, negatively impact the influencer’s credibility since trustworthiness is one
of the determining factors in source credibility.
The results did not provide significant evidence that the level of fashion interest impacted the
level of brand trust. The result can thus not support previous research on how
brand-consumer congruence impacts brand trust (Aaker 1999; Keller 1993; Leckie, Dwivedi
& Johnson 2022). This might be explained by the fact that the fit between an influencer and
the brand they collaborate with has a greater impact on brand trust, as the results suggest.
Thus, potentially outweighing how well consumers themselves resonate with the brand in
question.
5.2 Managerial Implications
Contrary to expectations derived from theories such as the Uncanny Valley Theory and
Anthropomorphism, the results did not show any significant impact of HVIs or AVIs on
brand trust. Therefore, we see no reason for marketers to exclude any of the influencer types,
either human, HVI, or AVI, due to the results of this research regarding anthropomorphism
and uncanniness. Instead, the results indicate that marketers should focus on developing high
authenticity and credibility of the influencers they collaborate with or design.
The findings, showing that consumers’ perceived authenticity (partially accepted H3) and
credibility of VIs mediate brand trust, provide marketers with guidance on how to use and
21
Märlegård & Raflund (2024)
design VIs in their influencer marketing strategies. As brand trust is positively influenced by
consumers’ perceptions of the authenticity and credibility of influencers, marketers should
either choose to collaborate with a VI that portrays genuineness, originality, and being true,
or design a VI with these characteristics in mind, which constitutes authenticity (Ilicic &
Webster 2016; Moulard, Garrity & Rice 2015). Marketers should choose to design or
collaborate with an HVI if their goal is to be perceived as authentic, as the results showed that
consumers perceive them to be more authentic than human influencers. As aforementioned,
this could be due to humans collaborating with brands to get an external reward, which VIs
do not. In contrast to an HVI, the results did not support H3 for an AVI. Hence, consumers’
perceived authenticity of an AVI did not mediate the level of brand trust, compared to the use
of a human influencer, indicating no difference between the two types. Also, there was no
significant difference in consumers’ perceived level of authenticity of an AVI and a human
influencer. Therefore, since perceived authenticity did not show significant importance for
AVIs, this would not be a crucial characteristic for marketers to focus on in the design or
portrayal of an AVI. Additionally, people perceive authenticity differently (Beverland &
Farrelly, 2010), which could make it better for marketers to instead focus on consumers’
perceived credibility of influencers, as that has a great impact on brand trust.
The results for credibility showed that consumers’ perceived credibility of a VI, either being
human-like or anime-like, significantly impacted the perceived credibility negatively,
compared to a human influencer. Hence, consumers do not find VIs credible, affecting the
acceptance of VIs, and thus, negatively affecting brand trust. This could stem from the fact
that VIs cannot actually try out a physical product, in this case, a new clothing collection. For
a VI to promote a physical product can negatively affect the trustworthiness of the VI,
contrary to if the VI would promote something virtual within its expertise (Lou et al. 2023).
This would, in turn, negatively impact the influencer’s credibility since trustworthiness is one
of the determining factors in source credibility. This makes it applicable for tech-oriented
marketers or brands that provide virtual products or services to use VIs who will be seen as
more credible due to their better fit with the products they promote. Therefore, marketers
should carefully consider and assess the fit between the influencer type and their target
audience’s preferences and ensure that their image aligns with the brand values and industry.
5.3 Limitations & Future Research
Although this research has provided interesting relationships and insights into VIs and brand
trust, it is not free from limitations. As the respondents of the treatment with the HVI could
not easily distinguish it from being a human influencer, an HVI, or an AVI, a review of the
design and features of the HVI would be appropriate in order to ensure an effective
manipulation. Since the study encountered the challenge of achieving a clear distinction
between a human influencer and an HVI, this could have impacted the validity of the
experimental design and results. Future research could explore alternative methods for
manipulating influencer types more effectively, such as improving the visual cues or the
narrative of the added text to enhance perceived differences between the influencers, for
22
Märlegård & Raflund (2024)
example, include a more detailed description of what the word “virtual” means in this
context. Some respondents could have interpreted it to refer to the virtual world of just the
digital online environment, when in fact it referred to an artificially created individual. This
could improve the participants’ perceptions of what constitutes a HVI.
The sample size of this research was quite small but enough to ensure statistical power.
However, issues can occur when having a small sample size, which can pose challenges to
the generalizability of the research findings and reduce external validity and statistical power.
Future research could include a broader demographic, as this study focussed on students.
Additionally, future research could also adopt an appropriate methodological strategy and a
larger sample size to enhance the robustness of their studies, by, for example, presenting the
survey to even more people, using random sampling or stratified sampling to enhance the
representatives of the sample and mitigate the effects of sampling bias. This could also
contribute to a greater understanding of consumers’ perceptions of VIs and their impact on
brand trust.
This study focused on the fashion industry which could have limited the generalizability of
the findings. An exploration of how influencer types affect brand trust in other industries and
settings could provide a more comprehensive understanding of its impact. As this research
examined brand trust stemming from consumers’ perceived authenticity and credibility of
influencers, future research could still include more factors that can affect the level of brand
trust. As previous research suggests, consumers trust brands to a greater extent if they are
familiar with the brand and have positive previous experiences with the brand in question
 (Akoglu & Özbek 2022; Dawar & Lei 2009; Ha, Pham & Le 2022).
This research explains 35% of brand trust, leaving 65% to be further explored. As previously
mentioned, brand trust is built of multiple components such as brand familiarity (Dawar &
Lei 2009; Ha, Pham & Le 2022), brand experience (Akoglu & Özbek 2022), and consumers’
emotional attachment (Bidmon 2017) to the brand in question. Incorporating these elements
into a similar study could enhance the explanatory power, providing a more comprehensive
understanding of what determines brand trust.
6. Conclusion & Contribution
This research provides insights into how consumers perceive different types of influencers
and guides marketers to recognize that while human influences may still hold a significant
level of trust among consumers, consumers’ perceptions of VIs vary and the effectiveness of
their portrayal can influence consumers’ trust in brands. Aiming to answer the first research
question, “How do different influencer types affect consumers’ trust in brands?”, the study
revealed that influencer types do not affect brand trust directly. This suggests that brands can
collaborate with whatever type of influencer they want, whether it is a human or virtual, and
instead prioritize building authentic and credible relationships with VIs and ensuring that
23
Märlegård & Raflund (2024)
their image aligns with the brand values and industry. This could be done by fostering
relationships with genuine and credible influencers, as these qualities impact consumers’ trust
in brands.
This leads us to answer the second research question, “How do the perceived authenticity and
credibility of virtual influencers affect brand trust?”. Consumers’ perceived authenticity and
credibility of influencers overall positively affect the level of brand trust, highlighting the
need for a brand to consider these qualities when collaborating with an influencer.
Consumers’ perceived authenticity of HVIs is, interestingly, higher than for a human
influencer, and is fully mediating the level of brand trust. This could indicate that consumers
perceive influencers as more authentic when their behaviors align with intrinsic motivations.
As some might perceive humans to be more prone to accepting external rewards, this could
be a reason why HVIs are deemed more authentic as they simply cannot earn any money
themselves. In contrast, consumers’ perceived authenticity of AVIs does not differ from
humans and does not mediate brand trust. Thus, authenticity is not a crucial characteristic for
AVIs and not something marketers should focus on for this type. However, consumers’
perceived credibility of VIs, regarding both types, is fully mediating the level of brand trust.
This shows the importance of aligning influencer characteristics with brand values and
industry to enhance consumers’ perceived credibility and trustworthiness of the influencer,
regardless of influencer type.
In conclusion, this study offers valuable insights into the complexity between influencer
types and brand trust in influencer marketing. The study uncovers the mediating effects of
consumers’ perceptions of influencer authenticity and credibility, contributes to a deeper
understanding of the dynamics around virtual influencers, and offers recommendations for
marketers.
24
Märlegård & Raflund (2024)
References
Aaker, J.L. (1999). The malleable self: The role of self-expression in persuasion. Journal of
marketing research, 36(1), pp. 45–57. doi:https://doi.org/10.2307/3151914.
Aaker, J., Fournier, S. & Brasel, S.A. (2004). When Good Brands Do Bad. The Journal of
consumer research, 31(1), pp. 1–16. doi:https://doi.org/10.1086/383419.
Ag, P. (2022). Virtual figures with real influence — Porsche & Virtual Influencers.
https://medium.com/next-level-german-engineering/virtual-figures-with-real-influence-493d1
5fbc370 [2024-01-05]
Akoglu, H.E. & Özbek, O. (2022). The effect of brand experiences on brand loyalty through
perceived quality and brand trust: a study on sports consumers. Asia Pacific Journal of
Marketing and Logistics, 34(10), pp. 2130–2148.
https://doi.org/10.1108/APJML-05-2021-0333
Allen, M. (2017). The SAGE Encyclopedia of Communication Research Methods. 1st edn.
Thousand Oaks, California: SAGE Publications, Incorporated.
doi:https://doi.org/10.4135/9781483381411.
Appel, G., Grewal, L., Hadi, R., & Stephen, A. T. (2020). The future of social media in
marketing. Journal of the Academy of Marketing Science, 48(1), pp. 79–95.
doi:https://doi.org/10.1007/s11747-019-00695-1
Arsenyan, J. & Mirowska, A. (2021). Almost human? A comparative case study on the social
media presence of virtual influencers. International Journal of Human-Computer Studies,
155(102694). doi:https://doi.org/10.1016/j.ijhcs.2021.102694
Audrezet, A., De Kerviler, G. & Moulard, J.G. (2020). Authenticity under threat: When
social media influencers need to go beyond self-presentation. Journal of Business Research,
117, pp. 557–569. doi:https://doi.org/10.1016/j.jbusres.2018.07.008
Belanche, D., Casaló, L.V., Flavián, M. & Ibáñez-Sánchez, S. (2021). Building influencers'
credibility on Instagram: Effects on followers’ attitudes and behavioral responses toward the
influencer. Journal of Retailing and Consumer Services, 61(102585).
doi:https://doi.org/10.1016/j.jretconser.2021.102585
Beverland, M.B. & Farrelly, F.J. (2010). The Quest for Authenticity in Consumption:
Consumers’ Purposive Choice of Authentic Cues to Shape Experienced Outcomes. The
Journal of Consumer Research, 36(5), pp. 838–856. https://doi.org/10.1086/615047
Bhandari, P. (2023). Control variables | What are they & Why do they matter?
https://www.scribbr.com/methodology/control-variable/ [2024-03-02]
25
Märlegård & Raflund (2024)
Bidmon, S. (2017). How does attachment style influence the brand attachment - brand trust
and brand loyalty chain in adolescents?. International Journal of Advertising, 36(1), pp.
164–189. doi:https://doi.org/10.1080/02650487.2016.1172404
Busser, J.A. & Shulga, L.V. (2019). Involvement in consumer-generated advertising.
International Journal of Contemporary Hospitality Management, 31(4), pp. 1763–1784.
doi:https://doi.org/10.1108/IJCHM-10-2017-0685
Chaudhuri, A. & Holbrook, M.B. (2001). The Chain of Effects from Brand Trust and Brand
Affect to Brand Performance: The Role of Brand Loyalty. Journal of Marketing, 65(2), pp.
81–93. doi:https://doi.org/10.1509/jmkg.65.2.81.18255
Chung, A. & Cho, H. (2017). Fostering Parasocial Relationships with Celebrities on Social
Media: Implications for Celebrity Endorsement. Psychology & Marketing, 34(4), pp.
481–495. doi:https://doi.org/10.1002/mar.21001
Dabiran, E., Farivar, S., Wang, F. & Grant, G. (2024). Virtually human: anthropomorphism in
virtual influencer marketing. Journal of Retailing and Consumer Services, 79.
doi:https://doi.org/10.1016/j.jretconser.2024.103797
Dawar, N. & Lei, J. (2009). Brand crises: The roles of brand familiarity and crisis relevance
in determining the impact on brand evaluations, Journal of business research, 62(4), pp.
509–516. doi:https://doi.org/10.1016/j.jbusres.2008.02.001.
Delgado-Ballester, E. (2004). Applicability of a brand trust scale across product categories.
European Journal of Marketing, 38(5/6), pp. 573–592.
doi:https://doi.org/10.1108/03090560410529222
Diamantopoulos, A., Sarstedt, M., Fuchs, C., Wilczynski, P. & Kaiser, S. (2012). Guidelines
for choosing between multi-item and single-item scales for construct measurement: a
predictive validity perspective. Journal of the Academy of Marketing Science, 40, pp.
434–449. doi:https://doi.org/10.1007/s11747-011-0300-3
Djafarova, E. & Rushworth, C. (2017) Exploring the credibility of online celebrities'
Instagram profiles in influencing the purchase decisions of young female users. Computers in
human behavior, 68, pp. 1–7. doi:https://doi.org/10.1016/j.chb.2016.11.009
Duffy, B., Smith, K., Terhanian, G. & Bremer, J. (2005). Comparing data from online and
face-to-face surveys. International Journal of Market Research, 47(6), pp. 615–630.
doi:https://doi.org/10.1177/147078530504700602
26
Märlegård & Raflund (2024)
Eggers, F., O’Dwyer, M., Kraus, S., Vallaster, C. & Güldenberg, S. (2013). The impact of
brand authenticity on brand trust and SME growth: A CEO perspective. Journal of World
Business, 48(3), pp. 340–348. doi:https://doi.org/10.1016/j.jwb.2012.07.018
Enkel, E. (2017). To get consumers to trust AI, show them its benefits.
https://hbr.org/2017/04/to-get-consumers-to-trust-ai-show-them-its-benefits [2024-01-22]
Garbin, L. (2022). What are Virtual Influencers? Knox Frost and 8 more cases.
https://boldcreators.club/knox-frost-virtual-influencers/ [2024-01-29]
Geuens, M. & De Pelsmacker, P. (2017). Planning and Conducting Experimental Advertising
Research and Questionnaire Design. Journal of Advertising, 46(1), pp. 83–100.
doi:https://doi.org/10.1080/00913367.2016.1225233
Giardino, J. (2021). Social Media Influencer Motivation: Exploring What Drives
Micro-Celebrities to Produce Content Using Social Exchange Theory. Master’s Thesis,
Communication and Media Technologies. Rochester Institute of Technology.
https://repository.rit.edu/cgi/viewcontent.cgi?article=11890&context=theses
Ha, Q.A., Pham, P.N.N. and Le, L.H. (2022) ‘What facilitate people to do charity? The
impact of brand anthropomorphism, brand familiarity and brand trust on charity support
intention’, International review on public and nonprofit marketing, 19(4), pp. 835–859.
Doi:https://doi.org/10.1007/s12208-021-00331-1
Hair Jr, J.F., Black, W.C., Babin, B.J. & Anderson, R.E. (2019). Multivariate Data Analysis.
8th edn. Pearson Education Inc.
Han, S. & Yang, H. (2018). Understanding adoption of intelligent personal assistants: A
parasocial relationship perspective. Industrial Management & Data Systems, 118(3), pp.
618–636. doi:https://doi.org/10.1108/IMDS-05-2017-0214
Hayes. (2014). Introduction to Mediation, Moderation, and Conditional Process Analysis: A
Regression-Based Approach (1st ed.). Guilford Publications.
Hoff, K.A. & Bashir, M. (2015). Trust in Automation. Human Factors, 57(3), pp. 407–434.
doi:https://doi.org/10.1177/0018720814547570
Huang, C. & Guo, R. (2021). The effect of a green brand story on perceived brand
authenticity and brand trust: the role of narrative rhetoric. The Journal of Brand
Management, 28, pp. 60–76. doi:https://doi.org/10.1057/s41262-020-00213-7
Hrastinski, S., Olofsson, A.D., Arkenback, C., Ekström, S. & Ericsson, E. (2019). Critical
Imaginaries and Reflections on Artificial Intelligence and Robots in Postdigital K-12
27
Märlegård & Raflund (2024)
Education. Postdigital Science And Education, 1(2), pp. 427–445.
doi:https://doi.org/10.1007/s42438-019-00046-x
Huang, M-H. & Rust, R.T. (2021). A strategic framework for artificial intelligence in
marketing. Journal of the Academy of Marketing Science, 49(1), pp. 30–50.
doi:https://doi.org/10.1007/s11747-020-00749-9
Hulland, J., Baumgartner, H. & Smith, K.M. (2018). Marketing survey research best
practices: evidence and recommendations from a review of JAMS articles. Journal of the
Academy of Marketing Science, 46(1), pp. 92–108.
doi:https://doi.org/10.1007/s11747-017-0532-y
Hussain, S., Melewar, T.C., Priporas, C-V., Foroudi, P. & Dennis, C. (2020). Examining the
effects of celebrity trust on advertising credibility, brand credibility and corporate credibility.
Journal of business research, 109, pp. 472–488.
doi:https://doi.org/10.1016/j.jbusres.2019.11.079.
Ilicic, & Webster, C. M. (2014). Investigating consumer–brand relational authenticity. The
Journal of Brand Management, 21(4), 342–363. https://doi.org/10.1057/bm.2014.11
Ilicic, J. & Webster, C.M. (2016). Being True to Oneself: Investigating Celebrity Brand
Authenticity. Psychology & Marketing, 33(6), pp. 410–420.
doi:https://doi.org/10.1002/mar.20887
Jhawar, A., Kumar, P. & Varshney, S. (2023). The emergence of virtual influencers: a shift in
the influencer marketing paradigm. Young Consumers, 24(4), pp. 468–484.
doi:https://doi.org/10.1108/YC-05-2022-1529
Keller, K.L. (1993). Conceptualizing, Measuring, and Managing Customer-Based Brand
Equity. Journal of marketing, 57(1), pp. 1–22.
doi:https://doi.org/10.1177/002224299305700101
Kim, S.Y., Schmitt, B. H. & Thalmann, N. M. (2019). Eliza in the uncanny valley:
anthropomorphizing consumer robots increases their perceived warmth but decreases liking.
Marketing Letters, 30, pp. 1–12. doi:https://doi.org/10.1007/s11002-019-09485-9
Kotzian, P., Stoeber, T., Hoos, F. & Weissenberger, B.E. (2020). To be or not to be in the
sample? On using manipulation checks in experimental accounting research. Accounting
Research Journal, 33(3), pp. 469–482. doi:https://doi.org/10.1108/ARJ-06-2019-0128
Krumpal, I. (2013). Determinants of Social Desirability Bias in SensitiveSurveys: A
Literature Review. Quality and Quantity, 47(4), pp. 2025–2047.
doi:https://doi.org/10.1007/s11135-011-9640-9
28
Märlegård & Raflund (2024)
Leckie, C., Dwivedi, A. & Johnson, L. (2022) ‘Examining drivers and outcomes of social
media brand engagement’, Asia Pacific journal of marketing and logistics, 34(2), pp.
350–367. doi:https://doi.org/10.1108/APJML-07-2020-0488
Lee, J.A. & Eastin, M.S. (2021). Perceived authenticity of social media influencers: scale
development and validation. Journal of Research in Interactive Marketing, 15(4), pp.
822–841. doi:https://doi.org/10.1108/JRIM-12-2020-0253
Lenth, R.V. (2001). Some Practical Guidelines for Effective Sample Size Determination. The
American Statistician, 55(3), pp. 187–193. doi:https://doi.org/10.1198/000313001317098149
Lim, R.E. & Lee, S.Y. (2023). “You are a virtual influencer!”: Understanding the impact of
origin disclosure and emotional narratives on parasocial relationships and virtual influencer
credibility. Computers in Human Behavior, 148(107897).
doi:https://doi.org/10.1016/j.chb.2023.107897
Lou, C., Kiew, S.T.J., Chen, T., Lee, T.Y.M., Ong, J.E.C. & Phua, Z. (2023). Authentically
Fake? How Consumers Respond to the Influence of Virtual Influencers. Journal of
Advertising, 52(4), pp. 540–557. doi:https://doi.org/10.1080/00913367.2022.2149641
Lou, C. & Yuan, S. (2019). Influencer Marketing: How Message Value and Credibility Affect
Consumer Trust of Branded Content on Social Media. Journal of Interactive Advertising,
19(1), pp. 58–73. doi:https://doi.org/10.1080/15252019.2018.1533501
MacKinnon, D.P. (2011). Integrating Mediators and Moderators in Research Design.
Research on Social Work Practice, 21(6), pp. 675–681.
doi:https://doi.org/10.1177/1049731511414148
Madsen, S. & Nielsen, L. (2010). Exploring Persona-Scenarios - Using Storytelling to Create
Design Ideas. Human Work Interaction Design: Usability in Social, Cultural and
Organizational Contexts, 316, pp. 57–66. doi:https://doi.org/10.1007/978-3-642-11762-6_5
McCracken, G. (1989). Who is the Celebrity Endorser? Cultural Foundations of the
Endorsement Process. The Journal of Consumer Research, 16(3), pp. 310–321.
doi:https://doi.org/10.1086/209217
McDermott, R. (2011). ‘Internal and External Validity’, in J.N. Druckman et al. (eds.)
Cambridge Handbook of Experimental Political Science. Cambridge University Press, pp.
27-40. doi:https://doi.org/10.1017/CBO9780511921452.003
Meyers, C.B. (2017). Social Media Influencers: A Lesson Plan for Teaching Digital
Advertising Media Literacy. Advertising & Society Quarterly, 18(2).
doi:https://doi.org/10.1353/asr.2017.0018
29
Märlegård & Raflund (2024)
Molenaar, K. (2023). Discover the top 12 virtual influencers for 2024 – listed and ranked!.
https://influencermarketinghub.com/virtual-influencers/#toc-10 [2024-01-26]
Molinillo, S., Ekinci, Y. & Japutra, A. (2019). A consumer-based brand performance model
for assessing brand success. International journal of market research, 61(1), pp. 93–110.
doi:https://doi.org/10.1177/1470785318762990
Morhart, F., Malär, L., Guèvremont, A., Girardin, F. & Grohmann, B. (2015). Brand
authenticity: An integrative framework and measurement scale. Journal of Consumer
Psychology, 25(2), pp. 200–218. doi:https://doi.org/10.1016/j.jcps.2014.11.006
Mutani, B. (2022). Virtual Influencers and their Social Media Appeal to Brands in the
Metaverse.
https://www.territory-influence.com/virtual-influencers-and-their-social-media-appeal-to-bra
nds-in-the-meteverse/ [2024-01-29]
Moulard, J.G., Garrity, C.P. & Rice, D.H. (2015). What Makes a Human Brand Authentic?
Identifying the Antecedents of Celebrity Authenticity. Psychology & Marketing, 32(2), pp.
173–186. doi:https://doi.org/10.1002/mar.20771
Nakada, M., Kavathatzopoulos, I. & Asai, R. (2021). Robots and AI Artifacts in Plural
Perspective(s) of Japan and the West: The Cultural–Ethical Traditions Behind People’s Views
on Robots and AI Artifacts in the Information Era. The review of socionetwork strategies,
15(1), pp. 143–168. doi:https://doi.org/10.1007/s12626-021-00067-8.
Ohanian, R. (1990). Construction and Validation of a Scale to Measure Celebrity Endorsers'
Perceived Expertise, Trustworthiness, and Attractiveness. Journal of Advertising, 19(3), pp.
39–52. doi:https://doi.org/10.1080/00913367.1990.10673191
Panjrath, Y. & Tiwari, S. (2021). “Why them, not me?”: A study exploring the impact of
following fashion influencers on Instagram on body image satisfaction of adolescent
girls and middle-aged women. International Journal of Psychosocial Rehabilitation, 25(02),
pp. 348–357. doi:10.37200/IJPR/V2512/PR320040
Pany, K. & Reckers, P.M.J. (1987). Within- vs. between-subjects experimental designs: a
study of demand effects. Journal of Practice and Theory, 7(9), pp. 39–53.
doi:https://web-p-ebscohost-com.ezproxy.ub.gu.se/ehost/pdfviewer/pdfviewer?vid=0&sid=a6
8b7d2b-c050-4faa-8eea-534b6967abd6%40redis
Porsche Newsroom. (2022). Virtual influencers in the automotive industry.
https://newsroom.porsche.com/en/2022/innovation/porsche-virtual-influencers-ayayi-automot
ive-industry-imaker-27480.html [2024-01-05]
30
Märlegård & Raflund (2024)
Portal, S., Abratt, R. & Bendixen, M. (2019). The role of brand authenticity in developing
brand trust. Journal of Strategic Marketing, 27(8), pp. 714–729.
doi:https://doi.org/10.1080/0965254X.2018.1466828
Presser, S. & Balir, J. (1994). Survey pretesting: Do different methods produce different
results. Sociological methodology, 24(5), pp. 73-104. doi:https://doi.org/10.2307/270979
Rajavi, K., Kushwaha, T. & Steenkamp, J-B.E.M. (2019). In Brands We Trust? A
Multicategory, Multicountry Investigation of Sensitivity of Consumers' Trust in Brands to
Marketing-Mix Activities. The Journal of Consumer Research, 46(4), pp. 651–670.
doi:https://doi.org/10.1093/jcr/ucz026
Rasmussen, M. (2021). 4 Reasons Brands Love to Work with Virtual Influencers.
https://www.virtualhumans.org/article/4-reasons-brands-love-to-work-with-virtual-influencer
s [2024-02-05]
Ryan, R.M. & Deci, E.L. (2000). Self-Determination Theory and the Facilitation of Intrinsic
Motivation, Social Development, and Well-Being. The American Psychologist, 55(1), pp.
68–78. doi:https://doi.org/10.1037/0003-066X.55.1.68
Sands, S., Campbell, C., Plangger, K. & Ferraro, C. (2022). Unreal influence: Leveraging AI
in influencer marketing. European Journal of Marketing, 56(6), pp. 1721-1747.
doi:https://doi-org.ezproxy.ub.gu.se/10.1108/EJM-12-2019-0949
Schallehn, M., Burmann, C. & Riley, N. (2014). Brand authenticity: model development and
empirical testing. The Journal of Product & Brand Management, 23(3), pp. 192–199.
doi:https://doi.org/10.1108/JPBM-06-2013-0339
Silva, M.J.B, Farias, S.A., Grigg, M.K. & Barbosa, M.L.A. (2020). Online Engagement and
the Role of Digital Influencers in Product Endorsement on Instagram. Journal of Relationship
Marketing (Binghamton, N.Y.), 19(2), p. 133–163.
doi:https://doi.org/10.1080/15332667.2019.1664872
Smith, E.R., Sherrin, S., Fraune, M.R., & Šabanović, S. (2020). Positive Emotions, More
Than Anxiety or Other Negative Emotions, Predict Willingness to Interact With Robots.
Personality & Social Psychology Bulletin, 46(8), pp. 1270–1283.
doi:https://doi.org/10.1177/0146167219900439
Stein, J-P., Breves, P.L. & Anders, N. (2022). Parasocial interactions with real and virtual
influencers: The role of perceived similarity and human-likeness. New Media & Society,
146144482211029. doi:https://doi.org/10.1177/14614448221102900
31
Märlegård & Raflund (2024)
The Influencer Marketing Factory. (2022). Virtual Influencers Survey + INFOGRAPHIC.
https://theinfluencermarketingfactory.com/virtual-influencers-survey-infographic/
[2024-02-01]
Thomas, V.L. & Fowler, K. (2021). Close Encounters of the AI Kind: Use of AI Influencers
As Brand Endorsers. Journal of Advertising, 50(1), pp. 11–25.
doi:https://doi.org/10.1080/00913367.2020.1810595
VirtualHumans. (n.d). Virtual influencers. https://www.virtualhumans.org/#influencers
[2024-02-03]
Viviani, M. & Pasi, G. (2017). Credibility in social media: opinions, news, and health
information—a survey. Wiley interdisciplinary reviews. Data mining and knowledge
discovery, 7(5), pp. 1209–n/a. doi:https://doi.org/10.1002/widm.1209.
Wang, C.L. (2021). New frontiers and future directions in interactive marketing: Inaugural
Editorial. Journal of Research in Interactive Marketing, 15(1), pp. 1–9.
doi;https://doi.org/10.1108/JRIM-03-2021-270
Weijters, B., Cabooter, E. & Schillewaert, N. (2010), The Effect of Rating Scale Format on
Response Styles: The Number of Response Categories and Response Category Labels.
International Journal of Research in Marketing, 27(3), pp. 236–47.
doi:https://doi.org/10.1016/j.ijresmar.2010.02.004
Wiedmann, K-P. & Von Mettenheim, W. (2021). Attractiveness, trustworthiness and expertise
– social influencers’ winning formula?. The journal of product & brand management, 30(5),
pp. 707–725. doi:https://doi.org/10.1108/JPBM-06-2019-2442.
Xin, Y., & Xiaogang, S. (2009). Linear regression analysis: Theory and computing. World
Scientific Publishing Company.
32
Märlegård & Raflund (2024)
Appendix
1. Manipulation 1: Human Influencer
Sarah is a social media influencer. She has been an influencer for 7 years and has
approximately 2.7 million followers on Instagram. She is known for her taste in fashion and
for being a top fashionista. Sarah posts content about her personal life and social good, as
well as promotional collaborations with many big multinational brands such as Nike, Dior,
Maybelline, etc. She has appeared in numerous headlines on TV, online, and in business
magazines, and is often referred to as one of the top influencers in fashion.
Sarah has recently signed a new contract to collaborate with another multinational fashion
brand to promote its products. She will be the face of the new collection and be present in
many marketing campaigns wearing the new products. She will also try them out and show
on her social media where her followers can participate in different social media activities
and ask questions.
2. Manipulation 2: Human-like Virtual Influencer
Sarah is a virtual social media influencer, meaning she is computer-generated and powered by
artificial intelligence. She has been an influencer for 7 years and has approximately 2.7
million followers on Instagram. She is known for her taste in fashion and for being a top
fashionista. Sarah posts content about her personal life and social good, as well as
promotional collaborations with many big multinational brands such as Nike, Dior,
Maybelline, etc. She has appeared in numerous headlines on TV, online, and in business
magazines, and is often referred to as one of the top influencers in fashion.
33
Märlegård & Raflund (2024)
Sarah, the virtual influencer, has recently signed a new contract to collaborate with another
multinational fashion brand to promote its products. She will be the face of the new collection
and be present in many marketing campaigns wearing the new products. She will also try
them out and show on her social media where her followers can participate in different social
media activities and ask questions.
3. Manipulation 3: Anime-Like Virtual Influencer
Sarah is an anime-like virtual social media influencer, meaning she is computer-generated
and powered by artificial intelligence. She has been an influencer for 7 years and has
approximately 2.7 million followers on Instagram. She is known for her taste in fashion and
for being a top fashionista. Sarah posts content about her personal life and social good, as
well as promotional collaborations with many big multinational brands such as Nike, Dior,
Maybelline, etc. She has appeared in numerous headlines on TV, online, and in business
magazines, and is often referred to as one of the top influencers in fashion.
Sarah, the virtual influencer, has recently signed a new contract to collaborate with another
multinational fashion brand to promote its products. She will be the face of the new collection
34
Märlegård & Raflund (2024)
and be present in many marketing campaigns wearing the new products. She will also try
them out and show on her social media where her followers can participate in different social
media activities and ask questions.
4. Cronbach’s Alpha: Authenticity
5. Cronbach’s Alpha: Credibility
6. Cronbach’s Alpha: Brand Trust
7. Manipulation check: ANOVA & Tukey’s Post Hoc
35
Märlegård & Raflund (2024)
8. Group Comparison on Brand Trust: ANOVA & Tukey’s Post
Hoc
36
Märlegård & Raflund (2024)
9. PROCESS MODEL 4 (without covariate)
37
Märlegård & Raflund (2024)
10. PROCESS MODEL 4 (with covariate)
38