“TikTok Made Me Buy It”

Influencer Marketing on TikTok: What Affects

Swedish Consumers' Purchase Intentions?

BACHELOR DEGREE PROJECT

THESIS WITHIN: Degree of Bachelor of
Science with a major in Business Administration
NUMBER OF CREDITS: 15
PROGRAMME OF STUDY: Marketing Management
AUTHORS: Johanna Nilsson, Filippa Nilsson, Ellen Johansson
Jönköping May 19th, 2023
Bachelor Thesis in Business Administration
Title: “TikTok Made Me Buy It”
Authors: Johanna Nilsson, Filippa Nilsson, Ellen Johansson
Tutor: Fatih Cura
Date: 2023-05-19
Key terms: Influencer marketing, Social media marketing, Electronic Word-of-Mouth,
Consumer behavior, Purchase intention, TikTok marketing.
_________________________________________________________________________

Abstract
This study investigates the impact of trustworthiness, perceived expertise, and parasocial
relationships on Swedish consumers' purchase intentions on TikTok, focusing on influencer
marketing. The problem addressed in this study is the lack of understanding of consumer
behavior on TikTok in the context of influencer marketing. While TikTok has become a
popular platform for influencer marketing due to its high engagement rates and entertaining
marketing concept, limited research exists on the factors that influence consumer purchase
intentions on the platform. The study aims to understand the factors that influence consumer
behavior on TikTok and its impact on purchase intentions, using the theory of persuasion. The
method used in this research is a quantitative study where the authors used a survey method to
collect primary data from 138 Swedish consumers. STATA was used to analyze the data, and
the findings reveal that the trustworthiness and parasocial relationships of influencers
significantly impact purchase intentions, while perceived expertise has no significant impact
on the platform TikTok. This study's results can guide marketing professionals and influencers
to develop effective influencer marketing strategies on TikTok, which offers immense
marketing opportunities for brands seeking to reach younger audiences with high engagement
rates.
Table of Contents
1. Introduction ................................................................................................................................................................... 1
1.1 Background.................................................................................................................................................... 1
1.2 Problem Discussion ....................................................................................................................................... 3
1.3 Purpose .......................................................................................................................................................... 5
1.4 Definitions ..................................................................................................................................................... 6

2. Theoretical Framework .............................................................................................................................................. 7

2.1 Social Media Marketing ................................................................................................................................ 7
2.1.1 Influencer Marketing ......................................................................................................................................................... 9
2.1.2 Electronic Word-of-Mouth (E-WOM) ..................................................................................................................... 10
2.1.3 Viral Marketing ................................................................................................................................................................ 11
2.2 Consumer Behavior ..................................................................................................................................... 12
2.3 Purchase Intention....................................................................................................................................... 13

3. Hypothesis .................................................................................................................................................................... 15
3.1 Theory of Persuasion ................................................................................................................................... 15
3.2 Perceived Characterizations ....................................................................................................................... 16
3.3 Characterizations ........................................................................................................................................ 16
3.3.1 Trustworthiness ................................................................................................................................................................ 17
3.3.2 Perceived Expertise ......................................................................................................................................................... 18
3.3.3 Parasocial Relationship.................................................................................................................................................. 18

4. Methodology ................................................................................................................................................................ 20
4.1 Research Design .......................................................................................................................................... 20
4.2 Data Collection Method .............................................................................................................................. 21
4.3 Data Analysis............................................................................................................................................... 22
4.3.1 Cronbach’s Alpha Coefficient ..................................................................................................................................... 24
4.3.2 Spearman’s Correlation ................................................................................................................................................. 25
4.3.3 Ordered Logistic Regression ....................................................................................................................................... 25
4.4 Methodological Limitations......................................................................................................................... 26
4.5 Reliability and Validity ................................................................................................................................ 27
4.6 Ethical Considerations ................................................................................................................................ 28

5. Result and Analysis .................................................................................................................................................... 31

5.1 Respondents Demographics ........................................................................................................................ 31
5.2 Usage of TikTok and Purchase Intention .................................................................................................... 32
5.3 Reliability Test ............................................................................................................................................. 34
5.4 Correlation Analysis .................................................................................................................................... 35
5.5 Regression Analysis ..................................................................................................................................... 38
6. Discussion ..................................................................................................................................................................... 40
6.1 Summary Results ......................................................................................................................................... 40
6.2 Hypothesis 1 ................................................................................................................................................ 41
6.3 Hypothesis 2 ................................................................................................................................................ 43
6.4 Hypothesis 3 ................................................................................................................................................ 46

7. Conclusion, Implication, and Limitations ............................................................................................................. 48

7.1 Conclusion ................................................................................................................................................... 48
7.2 Implications ................................................................................................................................................. 49
7.2.1 Theoretical Implications ................................................................................................................................................ 49
7.2.2 Practical Implications ..................................................................................................................................................... 49
7.3 Limitations and Future Research ................................................................................................................ 50

8. References .................................................................................................................................................................... 51

Appendix ........................................................................................................................................................................... 61
Appendix 1: The Rapid Rise of TikTok .............................................................................................................. 61
Appendix 2: Ordered Logistics Regression Odds Ratio .................................................................................... 61
Appendix 3: Mean Age of Respondents ............................................................................................................. 62
Appendix 4: Skewness and Kurtosis Tests for Normality .................................................................................. 62
Appendix 5: Survey Introduction Text ............................................................................................................... 63
Appendix 6: Survey Items .............................................................................................................................. 6464
Appendix 7: Variable Names ............................................................................................................................. 67
Appendix 8: Frequency and Percentage of Survey Answers ............................................................................. 70

Table of Figures
Figure 1. The Theory of Persuasion Adapted to Social Media Marketing. ...................................................... 15
Figure 2. Research Model ................................................................................................................................. 16

Table of Tables
Table 1: Variable Names and Connected Survey Question .............................................................................. 24
Table 2. Principles of Ethical Practices Adapted from Bell and Bryman (2007) ............................................. 29
Table 3. Respondents Demographic .................................................................................................................. 32
Table 4. Usage of TikTok and Purchase Intention ............................................................................................ 33
Table 5. Purpose of Use of TikTok .................................................................................................................... 34
Table 6. Preferred Type of Content ................................................................................................................... 34
Table 7. Cronbach’s Alpha Coefficient Analysis .............................................................................................. 35
Table 8. Spearman's Rho Correlation Coefficient ............................................................................................ 36
Table 9. Variance Inflation Factor Analysis ..................................................................................................... 37
Table 10. Ordered Logistic Regression Model.................................................................................................. 38
Table 11. Hypothesis Summary ......................................................................................................................... 41
1. Introduction

1.1 Background
“TikTok made me buy it”- a hashtag with 7.4 billion views on TikTok and endless videos
promoting products that the users discovered through the platform. Specific products receiving
major attention on TikTok are commonly occurring and leading to a significant increase in
demand, empty shelves in stores, and thousands of people signing up for waiting lists (Morris,
2021). The trend “TikTok made me buy it” is not only present on the platform, since brands
and retailers have acknowledged its major impact on consumer behavior as well. For example,
the makeup and skincare retailer Lookfantastic launched a beauty box in 2022 named “TikTok
made me buy it” including some of the most viral products on TikTok (Scrivener, 2022).
Influencer marketing on social media platforms has become a major business (Ye et al., 2021),
and TikTok is no exception. A shift in the most popular platforms to perform influencer
marketing on has been seen in recent years, with TikTok being perceived as a new beneficial
opportunity with high engagement among users (Haenlein et al., 2020).

TikTok was launched in 2016 and has since become the most popular platform for short-form
videos (Chu et al., 2022). The platform allows users to upload short videos which can be created
in the app using music, filters, and effects if wished. The content enables consumers to create,
upload, share, and engage with these short videos which can be up to 1 minute long when
creating the video in the app, while uploaded videos can be up to 3 minutes (TikTok, 2022). In
2021, TikTok was the number one most-downloaded app, with 656 million downloads
worldwide (Koetsier, 2021), and is also the platform that continues to rise most rapidly
(Buchholz, 2022). See Appendix 1. Moreover, a survey made by Svenskarna och Internet
(2022) regarding Swedish usage of TikTok showed that half of the population born in the 00s
use TikTok every day which also is the age group that most frequently uses TikTok. However,
the results also showed that the usage of people born in the 90s has increased significantly from
2021 to 2022 and a trend can be seen that it will continue to increase in usage within this age
group.

1
TikTok creators typically showcase themselves engaging in everyday tasks or comedic
reenactments of themselves, often within the confines of their own homes. This fosters a more
relaxed and personal relationship with their followers. Given TikTok's visual and entertaining
nature, content creators can effectively advertise products and services in an interactive and
captivating manner (Barta et al., 2023). In fact, Wiley (2022) suggests that the key to success
on the platform is to "Make TikToks, Not Ads," as viewers relish the experience of watching
one video after another. Furthermore, research has demonstrated that this type of content
resembles content generated by friends, allowing TikTokers to be perceived as relatable to the
viewer, thereby increasing their ability to influence (Barta et al., 2023).

Advertisers have always used celebrities and influential people to create engagement and
awareness of a brand because people tend to trust and look up to celebrities. Further, the term
influencer marketing is a relatively new term where the brands use an influencer on social
media to endorse their product or service which does not necessarily have to be a celebrity, tv-
star, etc. An influencer can be anyone on social media that builds an audience and have the
ability to influence that audience (Sammis et al., 2016). Furthermore, influencer marketing is
heavily relying on positive electronic word-of-mouth (e-WOM) created by the influencer about
the brand they are cooperating with. The impact mostly depends on the influencers´ believed
expertise, popularity, engagement with the audience, or reputation (Hudders et al., 2021). Since
the era of Social media began to be a part of our everyday life, influencer marketing has evolved
to become one of the most important marketing tools for brands (Sweeney et al., 2008;
Ismagilova et al., 2017).

Influencer marketing affects consumer behavior through its persuasion ability which
commonly creates purchase intention among the audience (Kanaveedu & Kalapurackal, 2022).
Purchase intention can be described as an individual’s conscious intention to proceed with a
purchase (Spears & Singh, 2004) and is commonly used as a tool to forecast future demand for
a product (Morwitz, 2014). Social media platforms, for example, TikTok, have become major
channels for marketing professionals aiming to create purchase intentions due to the presence
of influencer endorsement and electronic word-of-mouth on the platforms (Harrigan et al.,
2020).

2
1.2 Problem Discussion
Influencer marketing has become an effective strategy for social media marketing, particularly
on pioneering platforms such as Instagram. However, this success has extended to newer apps
like TikTok, where an increasing number of brands are partnering with TikTok creators. By
collaborating with popular creators, companies can better connect with younger audiences,
which has been a challenge for many businesses in the past (Dencheva, 2023a). According to
Haenlein et al. (2020), Facebook might be perceived as a suitable choice for influencer
marketing at first sight, due to its leading position as the platform with the most active users
and average time spent on the platform. However, from another perspective, it is possible to
make the prediction that Facebook’s most successful era has past and that the platform will
experience a decrease in active users in the future. Facebook is also possessing the position as
the platform with the lowest satisfaction rate, the highest rate of users considering leaving the
platform, and the only platform with a decrease in engagement among Instagram, YouTube,
Twitter, and TikTok.

Despite the business opportunities performing TikTok marketing, the research regarding the
topic is limited, possibly due to its relatively new position on the market and large increase in
popularity during recent years. In order for marketing professionals to take advantage of the
opportunities with TikTok, it is crucial to understand consumer behavior on the platform and
how consumers’ purchase intentions can be affected. According to Dencheva (2023b) is
TikTok the only platform compared to Instagram, YouTube, and Facebook, that is expected to
increase in the percentage of total influencer marketing expenditure between 2019 and 2024 in
the US. TikTok is forecasted to reach 18.5% of total influencer marketing expenditure in 2024.
The total expenditure for influencer marketing in Sweden during 2022 is expected to reach
1.464 billion SEK (Statista Research Department, 2022), thus it is possible to draw the
conclusion that major assets are, and will be, located toward influencer marketing on TikTok
which emphasizes the importance of extending the research available within the topic.

Influencer marketing on social media is a commonly researched topic, however, a large amount
of these studies are referring to more established platforms, for example, Instagram, YouTube,
and Facebook. TikTok has many characteristics in common with these platforms making the
existing research possible to apply on TikTok as well, however, Haenlein et al. (2020)
emphasize the importance of acknowledging each social media platform’s unique language,

3
culture, and demographics. Marketing strategies for the different platforms must be managed
individually in order to have maximum impact on consumer behavior, and companies aiming
for presence on social media should primarily focus on understanding the platform of choice’s
culture before implementing any marketing strategies to it.

Despite TikTok’s popularity, its future presence is discussed on a governmental level in

multiple nations, due to the platform being owned by the Chinese company ByteDance. Legal
forces within Europe and North America are expressing their concerns regarding Chinese
companies being required by law to provide their government with the personal data that the
companies possess. As a consequence, multiple nations have prohibited TikTok from being
downloaded on federal devices (Maheshwari & Holpuch, 2023). Additionally, The U.S. House
Foreign Affairs Committee voted in March of 2023 in President Biden’s favor allowing him to
move forward with the proposition regarding prohibiting TikTok from all devices within the
United States (Wilkie & Feiner, 2023). However, in Sweden, the resistance against the platform
is yet not as established, and multiple authorities are still allowed to have it downloaded on
their federal devices (Haglund, 2023). Haenlein et al. (2020) acknowledges the threat of
TikTok’s existence as well but claim that the importance of understanding the culture on the
platform still exists. An eventual prohibition for all United States citizens would only eliminate
the use of TikTok specifically, not the technology and functionality the platform is based on.
Thus, the concept of TikTok would in the case of prohibition be adopted by other actors on the
social media market and the knowledge of influencer marketing on TikTok would still be
applicable.

Altogether, increasing the research of influencer marketing on TikTok is necessary due to the
current research gap in the form of the absence of research within the area. As previously stated,
research focused on other platforms is not completely implementable on TikTok due to cultural
differences among the platforms. The market for influencer marketing on TikTok has
experienced significant growth in the recent years and is forecasted to continue on the same
path in the future. This emphasizes the need for research in order to guide companies toward
more accurate influencer marketing decisions leading to improved returns on the investments
originating from influencer marketing on TikTok. According to Haenlein et al. (2020), the
strength of TikTok as a platform for influencer marketing originates from its popularity among
the younger generation since they are more receptive to this form of communication. Additional
strengths are its high engagement among the users and TikTok’s concept of short videos

4
allowing marketing to be entertaining for the audience which increases the possibility of a
larger number of exposures. Furthermore, Wiley (2022) suggests that the key to success on
the platform is to "Make TikToks, Not Ads," as viewers relish the experience of watching the
entertaining content the platform provides. The strength of companies using influencers to
promote their brands is explained by Yesiloglu and Costello (2021). They describe that
communication that comes directly from the brand itself is less effective since it is still
perceived as an organization trying to sell its products. Influencers, on the other hand, are
perceived as more authentic and credible which creates an opportunity for a third-party
endorsement. In turn, they deliver engaging branded content and enhance communication with
the brand's target audience on social media platforms, which in this research is connected to
the platform TikTok.

Three commonly researched factors within influencer marketing are trustworthiness, perceived
expertise, and parasocial relationship (Onu et al., 2019; Aggad & Ahmad, 2021; Chung & Cho,
2017; Liu, 2022; Lou & Kim, 2019; Li & Peng, 2021; Wang et al., 2017; Bi & Zhang, 2022;
Reinikainen et al., 2020; Leite & Baptista, 2022). Masuda et al. (2022) connected these
variables with purchase intention by the theory of persuasion in order to research the extent to
which the variables impact purchase intention originating from influencer marketing on
YouTube. Due to the cultural differences between the platforms and the absence of research
regarding influencer marketing on TikTok, the theory of persuasion can be analyzed in the
context of TikTok as well in order to cover the existing research gap.

1.3 Purpose
This paper aims to investigate if the factors of trustworthiness, perceived expertise, and
parasocial relationship are affecting the purchase intention of Swedish TikTok users in the
context of influencer marketing on the platform, by using the theory of persuasion. Thus, the
purpose is to analyze the relationship between the factors trustworthiness, perceived expertise,
parasocial relationship, and purchase intention in order to describe the market of influencer
marketing on TikTok in Sweden. The purpose will be addressed by the following research
question:

RQ: Influencer Marketing on TikTok: What Affects Swedish Consumers' Purchase Intentions?

5
1.4 Definitions
TikTok: is a social media platform that allows users to upload short-form videos that can be
created directly in the app or uploaded to the app. The videos are between 1-3 minutes long
and contain creative content to engage the audience.
(Chu et al., 2022); (TikTok, 2022)

Electronic Word-of-Mouth (e-WOM): is where both positive and negative informal

statements are created online by consumers about a company or product, which is made
available to a large collection of viewers online.
(Mishra, 2018; Ismagilova et al., 2017)

Influencer: can be described as an individual who has the power to affect the purchase
decisions of others because of his/her authority, knowledge, position, or relationship with
his/her audience. An influencer does not necessarily have to be a celebrity, it can be a normal
person that creates content for a number of followers with the goal of building their name,
engagement, and online presence.
(Yesiloglu & Costello, 2021); (Geyser, 2023)

Purchase Intention: can be described as an individual’s conscious intention to proceed with a

purchase, where the intention is connected with the motivation of performing an action and
purchase entails that the motivation of the action is towards purchasing.
(Spears & Singh, 2004).

Theory of Persuasion: refers to the process of changing an individual’s attitude or behavior.

It suggests that persuasion can occur through two routes: the central route, where people
carefully analyze the message and its relevance to their values, and the peripheral route,
where people are influenced by superficial cues like the attractiveness of the persuader.
(Petty & Cacioppo, 1981)

Parasocial Relationship: is psychological association that users of media build with a

celebrity or media character, such as a social media influencer. These associations create a
feeling of intimacy with the media character as if they have a real relationship.
(Hwang and Zhang, 2018)

6
2. Theoretical Framework
The following literature has mainly been collected through the database Primo due to its large
collection of credible academic literature. Literature from Google Scholar has been reviewed
and included as well. The primary search, aimed at understanding the topics related to the
research question, generated a collection of 113 articles and books. However, additional
sources have been evaluated continuously during the construction of the theoretical framework
as new points of views was discovered. In summary, 38 sources were used in the theoretical
framework after a careful analysis regarding relevance and reliability. The framework consists
mainly of peer-reviewed articles and journals published between the years 2017-2023 to ensure
a contemporary perspective in the continuously changing world of marketing. However, some
older sources are included as well due to their reliability and richness. To support the included
peer-reviewed literature and to deepen the knowledge within the framework, additional books
have been included. To discover and investigate relevant literature, the following keywords
have been included in the search: “Influencer marketing”, “Social media”, “TikTok”, “Social
media influencer”, “Social media marketing”, “Electronic Word-of-Mouth”, “Consumer
Behavior”, “Purchase intention” and “Viral Marketing”. These keywords were chosen based
on the research question, objectives, hypotheses, and variables of interest. Moreover, these
keywords provide a framework for the study and guide the researcher's selection of appropriate
statistical analyses to analyze the data.

2.1 Social Media Marketing

Today social media has a significant role in our society. From the moment we wake up until
we go to bed, we are exposed to the digital society in some way throughout the day. People
from 31 countries answered a survey about their media presence and it showed that 96% of
people use their smartphone within an hour of waking up in the morning and often even before
they go up from bed (Tuten, 2021). Social media can be defined in many different ways,
however, according to Osborne‐Gowey (2014), “At its most basic level social media is simply
a collection of websites and applications designed to build and enhance online communities
for networking and sharing information”.

7
Further, social media marketing is rapidly growing and has changed the perception of
marketing (Tuten, 2021). Social media marketing has gone from being a single marketing tool
where firms could analyze and predict consumer behavior, to becoming a two-way
communication process between company and consumer. Companies want to create long-term
relationships through the use of social media marketing, even though there is no clear definition
of the framework of how to integrate social media with the companies’ marketing strategies
(Li & Peng, 2021). Moreover, consumers have gained more control in the way of
communicating directly with the brand where they in an online environment can comment,
collaborate and share their thoughts with both the brand and other consumers. This two-way
communication process has opened up the possibility for online opinion leaders, also called
“influencers”, where a strong online presence can affect consumers’ purchase decisions, both
directly and indirectly (Tuten, 2021).

Social media marketing is becoming one of the most common channels for advertising today
because of its possible range of viewers where companies can communicate with their
consumers, create word-of-mouth, collect data information, and manage customer service
relations, only to mention some of the benefits. The trend of using social media as a marketing
tool is expected to be an upward trend in the coming years both for business-to-business (B2B)
and business-to-consumer (B2C) (Tuten, 2021).

According to Koay et al. (2022), Social Media Influencers (SMI) can be described as social
media users that are famous for their presence in a niche area, with many followers where they
have the power to endorse and convince their followers in the decision-making process. An
SMI can be a celebrity that has the power to influence decisions, or an ordinary individual that
has created a big follower engagement. Hence, SMI can for example be everything from a
blogger, a traditional celebrity, an online entrepreneur, or a fashion enthusiast who advocates
their lifestyles and promotes the image of, for example, beauty brands.

8
2.1.1 Influencer Marketing
According to Influencer Marketing Hub, an influencer is: “…an individual who has the power
to affect the purchase decisions of others because of his/her authority, knowledge, position, or
relationship with his/her audience” (Geyser, 2023). An influencer does not necessarily have to
be a celebrity, it can be a normal person that creates content for a number of followers with the
goal of building their name, engagement, and online presence. Moreover, social media creates
an interactive environment where everyone on the platform has equal opportunities to create
and view content in any form which creates more equal power among people on the platform
(Yesiloglu & Costello, 2021).

Influencer marketing is a relatively new marketing tool that uses influencers on a social media
platform to communicate positive electronic word-of-mouth (e-WOM) on their channels. The
impact of the marketing mostly depends on the influencers’ expertise, popularity, or reputation
(Ye et al., 2021). Furthermore, WOM has long been recognized as one of the most crucial
communication channels among consumers, with the influence of opinion leaders being the
most effective aspect of this marketing strategy (Sweeney et al., 2008; Ismagilova et al., 2017).

According to Yesiloglu and Costello (2021), there are two ways of defining influencers´ role
in the marketing and advertising industry: influencer marketing and influencer relations.
Influencer marketing refers to marketing activities related to short-term and paid activities, and
influencer relations refers to activities that are unpaid and earned with media content that aims
for a long-term relationship with their audience. Furthermore, influencer marketing is a
common way to advertise today where marketing professionals and brands invest in the
influencers which in return creates content for the brands´ target audience. The brand itself also
uses communication on its platforms to communicate with its audience, however, Yesiloglu
and Costello claims that it is not as effective since it is still perceived as an organization trying
to sell its products. Influencers, on the other hand, are perceived as more authentic and credible
which creates an opportunity for a third-party endorsement where they can deliver engaging
branded content and enhance communication with the target audience on social media
platforms. Due to influencer marketing, brands can create closer relationships with their target
audience through influencers´ electronic word-of-mouth (E-WOM) and strong connection with
their followers that has the possibility to impact their follower's purchase decisions.

9
According to Reinikainen et al. (2020), parasocial relationships (PSR) are connected to the
term “influencer” since they can be described as imaginary relationships with media performers
that evolve over time to a believed deeper relationship only through the media consumption of
the performer. As the believed relationships become deeper, the person looking at the media
performer feels like they know the person as a friend. A media performer can trigger this
illusion through eye contact through the camera, the camera at eye level with the performer,
having private and informal conversations, greeting the audience, etc. Reinikainen et al.,
conducted a study to investigate if parasocial relationships, credibility, and the moderating role
of audience comments in influencer marketing, had a significant role in brand trust and
purchase intention. Their results showed that PSR with an influencer contributed to more trust
which further could reduce uncertainty people have towards a brand.

2.1.2 Electronic Word-of-Mouth (E-WOM)

Electronic word-of-mouth (E-WOM) originates from the concept of regular word-of-mouth
(WOM). The new definition was developed based on the new internet environment where both
positive and negative informal statements are created online by consumers about a company or
product, which is made available to a large collection of viewers online (Mishra, 2018;
Ismagilova et al., 2017). Further, one main difference from traditional WOM is that the
communication is perceived as non-commercial by the viewers. Moreover, one reason for this
can be that “influencers” or online “opinion leaders” can receive payment to share their
unbiased opinions about a product, service, brand, or company (Ismagilova et al., 2017).

E-WOM has become one of the most important marketing aspects today since it has a large
impact on both the consumer's pre- and post-purchase experience. Moreover, it has been
documented that e-WOM has a significant effect on the consumer's decision-making process.
Traditional communication media which includes television and magazines can be more costly
than marketing on different online platforms, which in turn has shown less effectiveness.
Furthermore, e-WOM is more beneficial in comparison to traditional marketing
communication, not only because it has significantly lower costs, but also has greater potential
to overcome consumer resistance to marketing messages (Zhou et al., 2019). E-WOM also
includes other online platforms like discussion forums, consumer review sites, blogs, shopping
websites, etc, where e-WOM is executed between anonymous users, but also includes social

10
media platforms. The quality of the information has great importance in the consumer's
evaluation process when it comes to both products and services (Erkan & Evans, 2018).

2.1.3 Viral Marketing

According to Petrescu (2014), Viral marketing is a concept described similar as to word-of-
mouth (WOM) but with enhancements towards the internet and social networks online. Viral
marketing is described to have a virus-like effect where the marketing information spreads like
a virus from person to person. This marketing tactic is based on the grounds that the person
“infected” by the marketing message willingly passes the message along to others by the means
of the internet, which could be everything from e-mail, videos, commercial newsletters, blogs,
or links to websites. A key point in viral marketing is to offer consumers incentives to pass
along the message, where the incentives often come from the entertainment potential but can
also come from financial benefits in the form of discounts or material incentives. Examples
have proven that the message needs to be innovative, distinct, creative, and generally different
from the other messages on the market to be able to generate consumer interest. Research has
shown that consumers acquired by viral marketing messages are proven to be more profitable
in the long run, compared to consumers acquired by traditional marketing. Further, the key
takeaway from viral marketing is that to create a buzz around the message, it must be
entertaining, creative, or shocking to be able to get consumers to pass the message along to
their friends and families.

E-WOM can be used in viral marketing as a type of marketing strategy, where firms can run a
controlled marketing campaign with the help of e-WOM and create a buzz around their brand
or product using opinion leaders. The distinct differences between e-WOM and viral marketing
are that e-WOM is described as unbiased and informal opinions created by consumers, whilst
viral marketing is more about creating a buzz with the help of e-WOM originally created by a
firm (Bao & Chang, 2016; Mishra, 2018). Moreover, viral marketing encourages consumers to
share information, whereas the “share” function on social media allows the user to forward a
message to a friend directly or through a direct link back to the app. By creating incentives
around the shared message, companies can increase product adoption and message forwarding
rate (Shao & Jing, 2022; Bhattacharya et al., 2019).

11
2.2 Consumer behavior
According to Foxall (2015), consumer behavior is the key knowledge for companies within a
competitive market in order to survive and grow their market share. The concept emphasizes
the knowledge about what, why, how, and where consumers make purchases and is the
underlying motivation behind marketing research. Behavior is commonly connected with
attitude, and their relationship can be presented as a two-way process where attitudes affect
behavior, and past behavior affects the present attitude. Product awareness can be viewed as
one of the founding pillars of consumer behavior. However, the determination to proceed with
a purchase is determined by the consumer’s motivation based on a large range of factors, for
example, needs, aspirations, and attitude. Furthermore, Foxall states that motivation is rarely
based on a single factor and researchers found it hard to measure them separately due to their
widely spread interrelationships. Another important factor within consumer behavior is
personality, which can be described as an individual’s motivational dispositions emerging from
the physical and social environment, and biological drives creating personality traits.
Marketing professionals are commonly considering the traits of their consumer segment when
developing marketing content, in order to attract them and affect their consumer behavior.

According to Solomon et al. (2019), consumer behavior was traditionally often referred to as
the purchase behavior of consumers and the interactions occurring during a purchase. However,
today marketing professionals have recognized consumer behavior as a continuous process
including all factors contributing to influencing the consumer before, during, and after a
transactional exchange. This theory also includes additional parties, for example, influencers,
impacting consumer behavior by sharing their opinions and recommendations. Furthermore,
Solomon et al. claim that influencers possess social power as they are having the capacity to
influence the behavior of their followers. The power commonly originates from three different
psychological factors. First, it can originate from referral power where the followers are
revering to the influencer and desiring to mirror the behavior of the influencer. Secondly,
information power where the influencer is perceived by the followers as obtaining valuable
information which the followers are acting accordingly to. The third factor is expert power
where the influencer is perceived as an expert by the followers who therefore trust the opinions
of the influencer. Additionally, the social power of influencers is providing opinion leadership,
which in an online context also can be referred to as power users, due to the followers valuing
the actions and opinions of the influencers and adapting their consumer behavior accordingly.

12
Research has indicated that the consumer behavior of young Swedish consumers, especially
females who commonly shop online and emphasize trust in the internet, is more likely to
include impulse buying tendencies. Additionally, their perception of risk within a purchase is
commonly uncertain regarding the need for the product (Ek et al., 2017). According to Hjort
and Lantz (2016), the consumer behavior of northern Europeans is affected by the return
policies within online retail shopping. The consumers are more likely to return products when
perceiving minor contradictions between their expectations and the actual product if returning
the product is free of charge. Offering free returns additionally increases the order frequency
while decreasing the average order value for northern European consumers.

According to Vrontis et al. (2021) the topic regarding how social media influencers are
affecting consumer behavior has become an inevitable topic within marketing research and
studies are indicating its impact on consumer behavior. However, the area of what is impacted
varies and can, for example, be purchase intention, brand attitude, and brand recall, among
many other areas within consumer behavior. Furthermore, the impact influencer marketing has
on consumer behavior may not always be positive, depending on the context and performance
of the marketing (Vrontis et al., 2021).

2.3 Purchase Intention

Purchase intention can be described as an individual’s conscious intention to proceed with a
purchase. Intentions are connected with the motivation towards performing an action, and in
research, it is commonly analyzed together with brand attitude, which is a perceptive summary
based on the individual’s evaluations. However, their relationship is not constant within the
researchers’ results (Spears & Singh, 2004).

Purchase intentions are commonly used within the marketing profession as a tool to forecast
future sales, measure the demand for newly launched products, and evaluate how actions may
affect consumers’ purchase behavior. Despite purchase intention being commonly used as
proxies for consumer behavior, their correlation is imperfect and consumers do not
uncommonly fail to act as they intended. For example, due to changed needs or economic
situation, or because of the marketing of a similar product being more appealing (Morwitz,
2014). In today’s digital society, social network platforms have become a major focus for

13
marketing professionals aiming to create purchase intention due to the networks enabling
electronic word-of-mouth and recommendations from influencers (Harrigan et al., 2020).
According to Laksamana (2018), multiple studies show that marketing on social media has a
major impact on purchase intention and that e-WOM interactions have a positive impact on
financial performance.

Commonly mentioned sources to increased purchase intention within the area of influencer
marketing is perceived trustworthiness (Onu et al., 2019; Aggad & Ahmad, 2021; Chung &
Cho, 2017; Liu, 2022; Lou & Kim, 2019; Masuda et al., 2022), perceived expertise (Liu, 2022;
Li & Peng, 2021; Masuda et al., 2022; Lou & Kim, 2019; Wang et al., 2017) and the parasocial
relationship that the audience has with the influencer (Lou & Kim, 2019; Bi & Zhang, 2022;
Reinikainen et al., 2020; Leite & Baptista, 2022; Masuda et al., 2022). Another well-researched
source to positively affects purchase intention is source credibility (Harrigan et al., 2020;
Magano et al., 2022; Liu, 2022; Masuda et al., 2022). Hovland et al. (1953) stated that
credibility is a two-dimension variable where expertise and trustworthiness are the factors
creating credibility, and this theory is still used within more modern research regarding
influencer marketing, for example, Schouten et al (2020) and Lou & Kim (2019).

14
3. Hypothesis
The hypotheses within the research will originate from the research regarding mediating roles
of characterizations within influencer marketing’s impact on purchase intentions made by
Masuda et al. (2022). The theory of persuasion is used to address the research gap and enable
research investigating if the included variables are affecting consumers’ purchase intention
when being exposed to influencer marketing on TikTok in regard to Swedish consumers.

3.1 Theory of Persuasion

According to Petty and Cacioppo (1981), does persuasion refer to the process of changing an
individual’s attitude or behavior. The elaboration likelihood model suggests that the process
can be categorized into two modes: central and peripheral. In the central route, elaboration is
more likely to take place, and the receiver focuses on value similarities between the message
and themselves. In the peripheral route, elaboration is less likely, and the receiver places more
emphasis on the attractiveness of the sender.

Masuda et al. (2022) connect the theory of persuasion and the elaboration likelihood model
with influencer marketing by claiming that consumers create characterizations of influencers
based on the influencers’ attributes. The perceived characterizations are then affecting the
consumers’ behavioral intentions, which in this research is purchase intention. See Figure 1.
The connection between the elaboration likelihood model and influencer marketing is
originating from research claiming individuals commonly have positive perceptions towards
strangers they are exposed to online, based on minimal and sometimes irrational indications
(Bacev-Giles & Haji, 2017). These questionable established perceptions can be connected with
the peripheral theory within the elaboration likelihood model due to its emphasis on
attractiveness (Masuda et al., 2022).

Figure 1. The Theory of Persuasion Adapted to Social Media Marketing.

15
However, as opposed to the research by Masuda et al. (2022), this research will exclusively
analyze perceived characterizations' impact on purchase intention due to their immediate
relationship. Influencer attributes are excluded since they are the source of the perceived
characterizations and therefore, do not have an immediate relationship with purchase intention.
This research is aiming to analyze what factors are affecting purchase intention, which
according to the theory of persuasion is primarily the perceived characterizations.

3.2 Perceived Characterizations

3.3 Characterizations
In the literature review, three fundamental factors were discovered contributing to the
characterization of social media influencers: trustworthiness, perceived expertise, and
parasocial relationship (PSR). Various studies (Masuda et al., 2022; Gomes et al., 2022; Lou
& Kim, 2019; Onu et al., 2019; Wang et al., 2017) have emphasized the significance of these
characterizations in influencing the purchase intentions of social media followers when
exposed to influencer marketing. Consequently, the research model presented in Figure 2 was
developed.

Figure 2. Research Model

16
3.3.1 Trustworthiness
Trustworthiness is related to how honest and caring the speaker is perceived to be by the
audience. If the source is seen as trustworthy, attractive, and as an expert, it can influence the
audience's attitude and behavior, including purchase intention (Masuda et al., 2020).
Trustworthiness is defined as the perceptions of honesty, integrity, and believability of an
endorser. Influencers tend to promote products in authentic, real-life settings, which can
increase perceptions of trustworthiness compared to celebrities (Schouten et al., 2020).
According to Ohanian (1990), trustworthiness is defined as the degree of confidence and
acceptance that the listener has in both the speaker and the message. Additionally, the impact
of source trustworthiness on the persuasibility of communication was investigated. The results
showed that a highly trustworthy communicator was more effective in producing attitude
change through opinionated messages. However, this relationship was not significant when
trustworthiness was low. Moreover, Chu et al. (2022) reported that social media influencers
and their followers connect based on trust and personality. The results revealed that social
media influencers’ trustworthiness significantly impacts consumers’ purchase intention.

According to Aggad and Ahmad (2021), influencer trustworthiness is the perception of

individuals based on sincerity, honesty, or truthfulness. While other factors have been
recognized by scholars, they emphasized that trustworthiness remains a critical factor for
influencers. The trustworthiness of influencers has a significant impact on consumers'
perceived credibility, information, purchase intention, e-WOM, and attitudes toward purchase
decisions. Numerous studies have found a strong positive relationship between social media
influencers’ trustworthiness and purchase intention in developed countries. Additionally,
influencers utilize various formats, such as videos, photos, and blogs, to develop strategies for
promoting products and services. The results of Aggad and Ahmad´s study also revealed that
the trustworthiness of social media influencers has a significant impact on consumers' purchase
intention. Thus, the following hypothesis is proposed:

H1: Trustworthiness is positively associated with followers’ purchase intention.

17
3.3.2 Perceived Expertise
Expertise refers to the extent of the endorser's perceived understanding, skills, and knowledge
(Hovland et al., 1953). In addition, Lou and Kim (2019) explain that source expertise refers to
the qualifications and knowledge of a source that enable them to form judgments about a
particular topic or subject.

Wang et al. (2017) further explain that the expertise of an endorser plays a crucial role in
persuading consumers to buy a product. Studies have shown that endorsers who are perceived
as more knowledgeable and skilled tend to be more persuasive and generate higher purchase
intentions. Furthermore, expertise has indicated to have a positive influence on both brand
attitude and purchase intention. In addition, it is found that expert celebrities are more effective
in increasing recall of product information than non-expert celebrities. Therefore, an endorser
who is perceived as highly competent and knowledgeable is generally considered more
persuasive than one with lower levels of perceived expertise.

The use of endorsements highlights the importance of influencers' expertise in influencing

consumers' purchase decisions. The endorser's proficiency, knowledge, and abilities enhance
the effectiveness of consumers' decision-making, leading to a greater likelihood of positive
purchase intentions (Liu, 2022). This is also supported by the research by Wang et al. (2017)
which describes that the expertise of an endorser refers to their qualifications, which can
significantly impact the level of conviction needed to persuade consumers to purchase a
product. Hence, the following hypothesis is proposed:

H2: Perceived expertise is positively associated with followers’ purchase intention.

3.3.3 Parasocial Relationship

Parasocial relationship (PSR) is according to Hwang and Zhang (2018) psychological
association that users of media build with a celebrity or media character, such as a social media
influencer. These associations create a feeling of intimacy with the media character as if they
have an authentic relationship. Moreover, the concept of parasocial relationships was according
to Reinikainen et al. (2020) initially introduced to describe the experience that audiences of
television, film, or radio may have with media performers, in which viewers develop a one-
sided relationship with the media personality. Chung and Cho (2017) describe that Parasocial

18
relationships develop when an individual is frequently exposed to a media figure, leading to a
sense of intimacy, perceived friendship, and identification with the media character.

PSR forms an intimate relationship between the user and media characters, which according to
Hwang and Zhang (2018) influences the user to believe the media character to be more
trustworthy as a source of information. In addition, PSR has therefore become an important
role in advertising effectiveness because of its persuasion power. Reinikainen et al. (2020)
further explain that when it comes to influencers, having a parasocial relationship can moderate
the impact of persuasion knowledge, meaning that audiences may be less concerned about
brand endorsements made by influencers they perceive as friends.

Moreover, Chung and Cho (2017) explain that consumers who developed parasocial
relationships with TV shopping hosts tended to be more satisfied with their shopping
experiences. Additionally, they discovered that consumers had a positive attitude toward
products and were more likely to intend to purchase them when endorsed by celebrities or
media figures with whom they had developed a parasocial relationship. Further, Lou and Kim
(2019) discovered that there is a positive connection between followers and their favorite
influencers, creating a parasocial relationship. This relationship is influenced by factors such
as the entertainment value of influencer-generated content, the expertise of influencers, their
trustworthiness, attractiveness, and the perceived similarity between followers and their
favorite influencers. Additionally, the perceived PSR was found to be associated with
materialistic views and purchase intentions. Thus, the following hypothesis is proposed:

H3: Parasocial relationship is positively associated with followers´ purchase intention.

19
4. Methodology
The methodology section of a research paper is a critical component that outlines the
procedures and techniques used to collect and analyze data. In quantitative research, this
section typically includes a detailed description of the research design, data collection methods,
and statistical analysis techniques employed to test hypotheses or explore relationships between
variables. A well-designed and rigorous methodology is essential to ensuring the validity and
reliability of the research findings. This section should provide a clear and comprehensive
account of the procedures used in the study to enable other researchers to replicate the study
and evaluate the credibility of the results (Collis & Hussey, 2014). In this chapter, the
methodology used to investigate the research question will be described to provide a detailed
account of the research design, data collection methods, and statistical analysis techniques
employed to address the research objectives.

4.1 Research Design

Collis and Hussey (2014) define a research paradigm as a philosophical framework that
provides guidance on how scientific research should be conducted. This framework is rooted
in the study of fundamental concepts, such as knowledge, reality, and existence, and is shaped
by researchers' beliefs and assumptions about the world. Essentially, a research paradigm
serves as a guiding framework that informs how research should be conducted based on a
researcher's philosophical perspectives. Positivism is a research paradigm that emphasizes the
use of objective and measurable data to develop and test hypotheses. It assumes that there is a
single, objective reality that can be observed and measured, and that knowledge can be gained
through empirical observation and experimentation. Thus, this paradigm is used for this
research.

The population of this study represents all Swedish TikTok users over 18 years old. This age
group was decided based on the legislation of the Swedish Code of Statutes (1949:381). The
legislation determines that minors are not allowed to make purchases or enter contracts online
without parental consent. Moreover, in this research, convenience sampling was applied.
Convenience sampling is a widely used non-probability sampling technique. It involves
selecting individuals who are readily available or accessible, such as those who are in close
proximity to a physical location or have easy access to the internet (Edgar & Manz, 2017).

20
However, a limitation of the convenience sampling method is that convenience sampling tends
to yield participants from a limited geographic area, with similar socioeconomic status or ethnic
backgrounds, which could potentially affect the study's findings. When participants are similar
in one or more aspects, it can lead to a biased outcome (Emerson, 2015). As for this study, the
respondents have been solicited through the author's social media platforms to reach a large
and accurate target group since social media comes with the benefits of being shareable by
others. Since the authors are originating from different areas within Sweden, the survey had
the possibility to reach a larger geographical spread. In addition, the knowledge and existence
of the survey have been spread by word-of-mouth (WOM).

4.2 Data Collection Method

The method used in this research is a quantitative study where the authors used a survey method
to collect primary data from Swedish consumers. Primary data refers to research data that
originates directly from the source, for example, through conducting one's own experiments,
conducting interviews, organizing focus groups, or as for this study, using questionnaire
surveys (Collis & Hussey, 2014). The measures utilized in the research model were adapted
from Masuda et al. (2022) to operationalize the variables. Certain terminologies were modified
in accordance with the focus of this research, for instance, "TikTokers" replacing "influencers".
The survey was designed using Google Forms and consisted of 23 questions. See Appendix 6.
The majority of the questions were closed-ended questions measured on a 5-point Likert scale.
The Likert scale is beneficial for measuring respondents' attitudes and opinions and is
commonly used in marketing research. The respondents indicate the extent to which they agree
or disagree on a rating scale typically ranging from 1 to 5 where 1 represents "strongly
disagree" and 5 represents "strongly agree” (Nemoto & Beglar, 2014). Additionally, the Likert
scale allows data to be transferred and analyzed using statistics, hence, appropriate for this
research. Further, the questionnaire was written in Swedish because the purpose of the study
only addresses Swedish consumers. Once the measurement instrument was created, a pilot test
was carried out to verify the sufficiency of the questionnaire. The pilot survey was sent out to
15 people within the target group. Once the pilot survey was collected, a new research question
was developed due to insufficient data in relation to the former research question. Furthermore,
the final survey was executed based on Masuda et al. (2022) to enable the inclusion of the
theory of persuasion and was also adapted to the platform TikTok. In addition, a control
question was added at the end of the survey to ensure that the respondent had understood what

21
the survey investigated and answered the questions intentionally. See Appendix 6. The data
was gathered between the period of the 24th of March and the 24th of April.

In this research, the focus is exploring how influencers' personal characteristics can influence
the likelihood of purchasing a product in the context of video advertising. Specifically, the
study investigates the role of TikTok, which has emerged as a popular platform for promoting
products through videos (Chu et al., 2022).

4.3 Data Analysis

When the data given from the survey was collected, an analysis of the data was constructed
using STATA. STATA is a general-purpose statistical software that is known for its wide range
of statistical routines, ease of data management, and custom publication-quality graphics
(Gutierrez, 2010). A comparative analysis between the perceived characterization of the
influencer and the purchase intention of consumers was executed as well as descriptive
statistics. In addition, a reliability analysis, a correlation analysis, and a regression model were
performed to analyze if there was any significance between the various characterizations and
purchase intention.

Since the data is measured on a 5-point Likert scale, the variables included in the data analysis
are of ordinal character. Collis and Hussey (2014) describe ordinal variables as numerical codes
allowing the data to be ranked in a specific order, for example, on a scale where 1 represents
low and 5 represents high. Furthermore, Collis and Hussey state that this category of variables
requires non-parametric analyzing methods which not rely on the assumption of normally
distributed data. Instead, non-parametric analyses arrange the frequency of the data in size
order and make calculations based on the ranks. However, due to that the ranks only are
approximates of the original data, the risk of achieving a Type II error is more likely than when
performing parametric tests. Type II error represents an incorrect rejection of the tested
hypothesis, due to insignificant values in the ranked data while the original data as a matter of
fact is significant and supports the hypothesis. Thus, non-parametric tests are commonly
described as less powerful and less reliable.

22
Whether the Likert scale should be analyzed with non-parametric tests as proposed by Collis
and Hussey (2014), or if it should be analyzed with parametric tests, has according to Harpe
(2015) been a subject of discussion since the scale was developed in the 1930s. Harpe proposes
that parametric analyzes can be used on Likert-scaled data where the obtained answers are
spread among at least five different response categories and the data is normally distributed.
Wadgave and Khairnar (2016) present the arguments from both of the perspectives within the
discussion. Those who argue for not performing parametric tests claim that means and standard
deviation are not possible to interpret for Likert-scaled data since it represents verbal statements
and that parametric tests are not suitable due to the common absence of normal distribution. In
contrast, those who argue for the use of parametric tests enhance its robustness when used on
composited scores and claim that large sample sizes commonly have a normal distribution with
makes ANOVA and t-tests applicable. Murray (2013) performed both parametric and non-
parametric correlation analysis on Likert-scaled data from the same sample and concluded no
significant difference between the results generated from the parametric Pearson’s correlation
and the non-parametric Spearman’s correlation.

The majority of the data within this paper fulfilled Harpe’s (2015) arguments for when a
parametric analysis is possible to be performed, because it possesses five different response
categories and since multiple variables have a normal distribution as presented in Appendix 4.
However, the analysis within this paper will follow the approach proposed by Collis and
Hussey (2014), that non-parametric tests should be performed on ordinal data due to two
reasons. First, the discussion of whether or not parametric tests are applicable to ordinal data
appears to mainly be a topic within journal articles. Since the articles do not appear to have
reached an official conclusion on the topic, this research will follow statements in academic
books claiming non-parametric tests as the most suitable method (Collis & Hussey, 2014;
Ryan, 2021; Harrison et al., 2021; Reinard, 2006). Secondly, since not every variable indicated
a normal distribution, variables would need to be excluded in order to do parametric tests which
could make the analysis of the factors affecting purchase intention less rich.

In order to make the analysis in STATA more efficient and easier to interpret, the variables
were coded into shorter names as described in Appendix 7. Table 1 presents the coding of the
variables included in the correlation and the regression analysis, in order to make the
interpretation of the analysis more convenient.

23
 Name Question

T1 I get inspired by TikTokers I believe are honest

T2 I get inspired by TikTokers I believe are reliable

T3 I get inspired by TikTokers that are acting according to their beliefs and values

E1 I get inspired by TikTokers that are proficient

E2 I get inspired by TikTokers that I believe are knowledgeable in the area of

what they are recommending

E3 I get inspired by TikTokers I believe are experts in a special area

PR1 I get inspired by TikTokers which makes me feel comfortable as if it was a

close friend

PR2 I get inspired by TikTokers that I believe are natural and down-to-earth

PR3 I get inspired by TikTokers that understand more about areas that I want to
learn more about

PR4 I get inspired by TikTokers whom i would like to meet in real life

PR5 I get inspired by TikTokers that I find attractive

Table 1: Variable Names and Connected Survey Question

4.3.1 Cronbach’s Alpha Coefficient

According to Collis and Hussey (2014), Cronbach’s alpha coefficient is commonly used within
research to analyze the internal reliability within multiple-item scales. When performing the
test, each included item is correlated individually with the other items related to the sample
construct. Consequently, an alpha coefficient is presented representing the reliability based on
the average inter-item correlation. The resulting coefficient ranges from 0 to 1, with higher
values indicating greater internal consistency. A Cronbach’s alpha of 0.8 and above is generally
considered reliable according to Collis and Hussey (2014), however, other authors claim that
an alpha of 0.7 and above is indicating reliability as well (Harrison et al., 2021).

24
4.3.2 Spearman’s Correlation
Spearman’s correlation analysis, also known as Spearman’s rho, is a non-parametric analysis
measuring the linear association between the included variables. The test ranges the data values
in size order and analyses the ranking differences between the variables. The presented
correlation coefficients from the test represent the extent of linear dependence between the
variables, where 0 represents no linear association, -1 perfect negative linear association, and
1 perfect positive linear association. However, when interpreting the correlation between two
variables it is important to acknowledge that correlation is not evidence of causality and that
the correlation may originate from the variables’ association with a third variable (Collis &
Hussey, 2014). For example, there may exist a correlation between ice cream sales and number
of drownings, since both situations typically occur during high temperatures.

4.3.3 Ordered Logistic Regression

According to Collis & Hussey (2014), when aiming to predict the outcome from non-
parametric variables, logistic regression is preferred. Since the dependent variable within this
research is of ordinal character which needs to be considered in the regression analysis, the
ordered logistic regression model will be adopted based on literature claiming its suitability for
regression analyses of ordinal data (Hilbe, 2009; McNulty, 2021; Osborne, 2015; O’Connell,
2006).

According to Hilbe (2009) is the ordered logistic regression model a non-parametric analysis
originating from the binary response model allowing analyzes of binary variables. Binary
variables only consist of two response options, commonly 1 and 0 to identify absence and
presence, which the linear probability regression model does not have the ability to efficiently
analyze. The ordered logistic regression model can be used to analyze data with ordered
responses, for example, binary data, however, the model is most frequently applied to data with
three or more ordered response levels. The ordered logistic regression model applies the
assumption of all the response levels possessing equal coefficients, thus the model is known as
the proportional odds model as well.

25
McNulty (2021) emphasizes the ordered logistic regression model’s suitability for Likert scale
data and describes further the models' possibility of accounting for the stepwise characteristic
that Likert data possesses. The difference between 1 and 2 on the Likert scale is necessarily not
equal to the difference between 2-3 in a psychological context, thus it can not be treated as
continuous and linear.

4.4 Methodological Limitations

Quantitative research involves the systematic collection, analysis, and interpretation of
numerical data to investigate relationships, patterns, and trends. While quantitative research
can provide valuable insights, it is not without its methodological limitations. Methodological
limitations in quantitative research refer to the potential constraints, weaknesses, or biases that
may affect the validity, reliability, or generalizability of the findings. It is crucial for researchers
to identify and acknowledge these limitations in their research reports to ensure transparency
and credibility (Collis & Hussey, 2014). In this section, the methodological limitations
encountered in this quantitative research will be discussed, including issues related to sampling,
measurement, and statistical analysis. By critically examining and addressing these limitations,
the aim is to provide a balanced and comprehensive interpretation of our research findings.

Due to this study’s research topic being relatively new in the literature, there is a lack of
previous studies within TikTok marketing. However, the existing literature within the topics
related to the research question is up-to-date and relevant to this study. A further limitation was
the sample size and population of the survey. The sample size is a methodological limitation
because it affects the generalizability of the findings. A larger sample size tends to provide
more representative results of the population being studied. If the sample size is too small, the
findings may not accurately reflect the attitudes or opinions of the population being studied. In
addition, small sample sizes can lead to lower statistical power, which can decrease the chances
of finding a significant effect even if one exists in the population. However, based on the time
constraint of one month, the sample size of 138 is useful in this study. Additionally, Vittinghoff
and McCulloch (2007) confirm this by explaining the commonly accepted "rule of thumb" in
statistical analysis that logistic and Cox models should be used with a minimum of 10 outcome
events per predictor variable (EPV).

26
The target audience in this research is individuals over 18 years old, however, heavy users of
TikTok are commonly younger. Hence, a large age group is missing in this research but due to
the legislation regarding online purchasing in Sweden, the chosen age group was the only
relevant one for this study. Although the Likert scale is a valuable measure tool in quantitative
research and when using a survey, it comes with some limitations. Since the scale contains
limited answers between 1-5 in this study, it may not capture the full range of opinions or
attitudes that participants may hold. This can lead to a loss of nuance in the data and potentially
inaccurate conclusions. Moreover, the interpretation of scale points can be subjective and may
vary among individuals. For example, what one person considers to be a "strongly agree"
response may be different from what another person considers to be a "strongly agree"
response. Additionally, participants may feel pressure to respond in a socially desirable way,
rather than answering honestly. This can lead to biased data and inaccurate conclusions.

4.5 Reliability and Validity

Collis and Hussey (2014) define reliability in the context of research as the extent to which the
results obtained from a research study are consistent and repeatable over time and under
different conditions. They state that reliability is a necessary condition for validity, which is
the degree to which research findings accurately reflect the reality being studied. Reliability
can according to Collis and Hussey (2014) be assessed using various techniques. Since this
research aims to analyze data using statistics, the widely used test, Cronbach's alpha coefficient
was used to ensure the reliability of the measures.

Cronbach’s Alpha Coefficient is a statistical measure of internal consistency reliability.

Internal consistency reliability refers to the extent to which different items in a measure or scale
are measuring the same underlying construct (Collis & Hussey, 2014).

Further, the validity of the research has been ensured by only using data from participants that
are Swedish consumers and users of TikTok which the participants were informed of before
answering the survey, see Appendix 5. Additionally, a control question was added at the end
of the survey to ensure that the respondents have been paying attention to the questions and
provided accurate responses. Hence, by including a control question the authors can assess the
validity of the responses to the other questions in the survey. If respondents are not paying

27
attention or are providing inaccurate responses, the control question identifies those individuals
and therefore, they can be excluded from the analysis.

4.6 Ethical Considerations

According to Collis and Hussey (2014), the ethics of research involve the way in which the
research is carried out and how the outcomes or discoveries are communicated. To ensure the
ethical consideration in this study, the authors have adhered to the ethical guideline by Bell and
Bryman (2007). Bell and Bryman established a guideline consisting of 11 principles for
conducting management research. These principles are presented in Table 2 followed by
descriptions on how the authors have implemented these in the research. The first paragraph
summarizes the connection between the principles 1-6 and the second paragraph 7-11.

Ethical Principle Definition

1) Harm to participants The potential to cause harm through the

research process and the need to ensure
physical and psychological well-being either
of research participants, the researcher, or
others.

2) Dignity The requirement to respect the dignity of

research participants, researchers, or others
and avoid causing discomfort or anxiety.

3) Informed consent The need to ensure the fully informed

consent of research participants.

4) Privacy The need to protect privacy of research

subjects or avoid invasions of privacy.

5) Confidentiality The requirement to ensure confidentiality of

research data whether relating to
individuals, groups, or organizations.

28
 6) Anonymity The protection of anonymity of individuals
or organizations.

7) Deception The potential for deception through the

research process, either through lies or
behavior that is misleading.

8) Affiliation The need to declare any professional or

personal affiliations that may have
influenced the research, including conflicts
of interest and sponsorship, including
information about where funding for the
research has come from.

9) Honesty and Transparency The need for openness and honesty in

communicating information about the
research to all interested parties, including
the need for trust.

10) Reciprocity The idea that the research should be of

mutual benefit to researcher and participants
or that some form of collaboration or active
participation should be involved.

11) Misrepresentation The need to avoid misleading,

misunderstanding, misrepresenting, or false
reporting of research findings.

Table 2. Principles of Ethical Practices Adapted from Bell and Bryman (2007)

These ethical principles were considered during the whole process. The research design and
topic were conducted to ensure that no harm to potential participants was at risk. The
participants were provided with pertinent information in the survey introduction, including an
overview of the research purpose, the number of survey items, an estimated response time,
GDPR regulations, and the author's contact details. Furthermore, to ensure a comprehensive
understanding of the survey items, participants were informed about the theoretical

29
explanations of the three perceived characterizations at the beginning of each chapter. This
approach aimed to facilitate participant comprehension and promote accurate responses. The
questions are professionally formulated and aim to only address relevant aspects of this study,
thereby minimizing the risk of causing any discomfort or anxiety among participants.
Additionally, the participation was optional and all the respondents were anonymous. Hence,
the ethical principles (1-6) are acknowledged.

To further ensure the ethical considerations of principles 7-11, the authors provided full
disclosure of the purpose and methodology of the study to the participants as well as ensured
that the participants understand their role and the possible outcomes of the study before they
agree to participate. See Appendix 5. Moreover, by conducting the research independently from
any company, we have minimized the potential for ethical concerns related to affiliations that
could influence the research. Additionally, as there were only three researchers involved, any
personal affiliations were limited. Further, to ensure honesty and transparency, the raw data are
available upon request by still ensuring the anonymity and privacy of our research participants,
and providing clear and accurate reporting of the findings from the analysis. Finally, by
conducting an inferential statistical analysis of empirical data, we aim to reduce the chances of
reporting false findings. Additionally, we strive to minimize the risk of any misunderstandings
through a transparent research process and clear research objectives. misrepresentation by
ensuring that the study is designed and conducted in an unbiased and objective manner.

30
5. Result and Analysis
In this chapter, data has been collected from a sample size of 138 Swedish consumers to
instigate purchase intention on TikTok in regard to the factors of trustworthiness, perceived
expertise, and parasocial relationship. The data presented in this report are derived from a
survey conducted with a sample of 142 participants, where only 138 were useful. One response
was excluded due to the age constraint of 18+ and three responses were excluded from the
analysis as these participants indicated a lack of understanding about the purpose of the survey
when answering the control question.

The survey questions have been in regard to the context of influencer marketing on the platform
by using the theory of persuasion. Further, the authors have analyzed the data in the form of
descriptive statistics, Cronbach’s alpha, Spearman’s correlation, and ordered logistic
regression. The analysis is focused on the factors of trustworthiness, perceived expertise,
parasocial relationship, and purchase intention in order to describe the market of influencer
marketing on TikTok in Sweden.

RQ: Influencer marketing on TikTok: What affects Swedish consumers' purchase intentions?

5.1 Respondents Demographics

Data analysis plays a crucial role in understanding and interpreting research findings in various
disciplines. Descriptive statistics are summarized data that are presented in a clear and
comprehensible manner. The data may be graphical, numerical, or tabular (Anderson et al.,
2014). In this study, descriptive statistics are employed to examine the demographic
characteristics of the sample, summarize the key variables of interest, and provide a clear
picture of the collected data. The respondent demographic data are presented in Table 3.

As presented in Table 3, 70,29% were female and 29,71% were male. The results convey that
the mean age of the sample population is 24 years. See Appendix 3. The majority of
respondents were aged between 23 and 26 years (61,59%), followed by 18 to 22 years
(25,39%), and 27 to 56 years (13%). Furthermore, the results from the descriptive statistics
indicate that the majority of the participants use TikTok multiple times a day, (60,87%),

31
followed by respondents answering that they are using the platform once a day (16,67%). Only
8,7% indicated that they use TikTok less often than a few times a week. See Table 3.

Table 3. Respondents Demographic

5.2 Usage of TikTok and Purchase Intention

In this section, descriptive data about TikTok usage and purchase intention will be presented.
The response alternatives ranged from 1-5, where 1 indicated that the participants “strongly
disagree” with the statement, and 5 indicated that they “strongly agree”. In addition, indicators
4 and 5 would consider that the participants agree with the statement presented (Nemoto &
Beglar, 2014). The results from the data demonstrate that 82 participants out of 138, which
equals 59,42% would agree that they have experienced a purchase intention from content by
TikTok. Moreover, only 21,02% answered that they disagreed with the statement of purchase
intention on TikTok. The table also shows that 61,59% of the participants get inspired by the
content they see on TikTok, whereas 30,43% of the participants have actually purchased a
product or service they saw on TikTok. The numbers also indicate that 47,83% believe that

32
recommended items are more trustworthy than traditional commercials whereas 34,06%
disagree with the statement. See Table 4.

Table 4. Usage of TikTok and Purchase Intention

A question was utilized for the understanding of why the participants use TikTok, where it was
possible to fill in more than one answer. In Table 5, the results convey that the highest purpose
of using TikTok with 94,93%, is entertainment whereas inspiration is the second highest
purpose with 50,72%. Product recommendation came in third place with 38,41%.

33
 Table 5. Purpose of Use of TikTok

Table 6 represents the respondents preferred types of content on TikTok where it was possible
to fill in more than one answer. Further, the most preferred content was entertainment with
81,16%, in second place was recipes with a rate of 70,29%, and in third place was
DIY/rebuilding with 60,87% of the answer rate.

Table 6. Preferred Type of Content

5.3 Reliability Test

To ensure the reliability of the hypothetical constructed variables within this research,
Cronbach’s alpha coefficient analysis was conducted to test the variables’ internal consistency
and reliability. The alpha represents the average inter-item correlation which is calculated by
correlating the items individually with the other items related to the sample construct. A
coefficient above 0.8 indicates reliability for the variables (Collis & Hussey, 2014).

34
 Table 7. Cronbach’s Alpha Coefficient Analysis

According to Table 7 the Cronbach's alpha for the variables included in this research is above
0.8 and they are considered reliable for further analysis.

5.4 Correlation Analysis

According to Collis and Hussey (2014), the Spearman's rho correlation coefficient is suitable
for non-parametric data measured on an ordinal scale, for example, the Likert scale. It accounts
for the sizes of the data values and analyzes the differences between the rankings of the
variables. Correlations between 0.00 and 0.39 are considered a low positive correlation, 0.4 to
0.69 a medium positive correlation, 0.7 to 0.89 a high positive correlation, and 0.9 to 0.99 a
very high positive correlation. The limit for significance within business and management
research is commonly 0.05, meaning a 5% risk of the results being due to chance is accepted.
The survey question representing the dependent variable Purchase_intention is based on the
question “I feel that I want to purchase products/services I see on TikTok”. The labels for the
independent variables represent the areas of research within the survey where T represents
trustworthiness, E represents expertise, and PR represents parasocial relationships. The number
following the letter represents the specific question connected to the variable within the area of
focus, for example, T1 represents the first question within the trustworthiness section. See
Table 1 for the exact coding structure.

35
 Table 8. Spearman's Rho Correlation Coefficient

Based on Table 8, all correlations are significant and positive, however, no correlation between
the dependent variable and the independent variables is high or very high.

Within the area regarding trustworthiness, the variable T1 representing honesty has a medium
positive correlation with purchase intention (r = 0.5397, p = 0.0000) as well as T2 (r = 0.5073,
p = 0.0000) representing reliability. T3 representing acting according to your beliefs and values
has a low positive correlation with purchase intention (r = 0.1820, p = 0.0327). The variable
E1 representing proficiency within the area of expertise has a medium positive correlation with
purchase intention (r = 0.4436, p = 0.0000), while E2 (r = 0.3467, p = 0.000) representing
knowledgeability and E3 (r = 0.3013, p = 0.0003) both have a low positive correlation with
purchase intention. Regarding the section for a parasocial relationship, PR2 has a medium
positive correlation with purchase intention (r = 0.4404, p = 0.0000) and the remaining
variables PR1 (r = 0.3896, p = 0.0000), PR3 (r = 0.3238, p = 0.0001), PR4 (r = 0.2745,

36
p = 0.0011) and PR5 (r = 0.3849, p = 0.0000) have a low positive correlation with purchase
intention.

Furthermore, based on Table 8 the independent variables T1 and T2 have a high positive
correlation (r = 0.7526, p = 0.0000) which indicates potential multicollinearity. According to
Collis and Hussey (2014), multicollinearity may occur when the correlation between two
variables is higher than 0.7, which could interfere with the reliability of multiple regression
analysis, cause unreliable standard error estimations, and difficulty to establish the effect of
individual independent variables. Thus, an variance inflation factor (VIF) analysis was
conducted to identify the level of multicollinearity and ensure no multicollinearity will interfere
with the regression analysis.

Table 9. Variance Inflation Factor Analysis

According to Alin (2010), if a VIF is larger than 10 it can be considered to have a strong
multicollinearity, which should be avoided within a regression analysis in order to ensure the
result is reliable. Based on Table 9 no variable exceeds the VIF limit of 10 and all independent
variables can be included in the regression analysis.

37
5.5 Regression Analysis
The regression analysis was conducted by the ordered logistic regression model which
according to Hilbe (2009) is suitable for ordinal data and assuming proportional odds where
the coefficients are assumed to be constant for each level of possible response. An ordered
logistics regression was conducted together with the model command performing a likelihood
ratio test to ensure proportional odds. In order to proceed with the analysis of the regression
the likelihood ratio test should be non-significant indicating the existence of proportional odds.
Based on Table 10 is the result from the likelihood ratio test non-significant (p>0.05), which
proves proportional odds among the variables and ensures further analysis of the regression as
reliable.

Table 10. Ordered Logistic Regression Model

38
Based on Table 10 the regression model is significant (p = 0.0000) indicating it is a good fit
for the data. T1 is a significant (p = 0.002) positive predictor of purchase intention where an
increase of one unit of perceived honesty increases purchase intention by 0.823 units, thus it is
possible to accept H1 stating that trustworthiness is positively associated with purchase
intention. Furthermore, PR2 is a significant (p = 0.028) predictor of purchase intention by
increasing purchase intention with 0.507 units for each additional unit of perceived “down-to-
earth” characterization. The final significant variable is PR5 (p = 0.012) which will increase
purchase intention by 0.413 units when increasing the perceived attractiveness with one unit.
Due to PR2 and PR5 being significant and positive is it possible to accept H3 as well indicating
that parasocial relationship is positively associated with purchase intention. None of the
variables within the area of expertise are a significant predictor of purchase intention, thus H2
will be rejected and expertise is not proven to be positively associated with purchase intention.
Additionally, odds ratios were obtained for the regression model demonstrated in Appendix 2,
where T1 obtained an odds ratio of 2.277, PR2 obtained 1.660, and PR5 obtained 1.511.

39
6. Discussion
TikTok has emerged as the most popular platform for short-form videos since its launch in
2016. Its immense popularity is evident from the 656 million downloads it received worldwide
in 2021 (Koetsier, 2021). TikTok has become a preferred platform for marketing professionals
to create engagement and awareness of their brand, and influencer marketing has played a
crucial role in this. Social media platforms like TikTok have become vital channels for
marketing professionals to create purchase intention among the audience. TikTok's unique
concept of short videos allows marketing to be more entertaining and interactive, thereby
increasing its effectiveness. Moreover, TikTok's popularity among the younger generation and
its high user engagement provide significant business opportunities for companies seeking to
promote their brands. Thus, TikTok holds immense potential for businesses looking to improve
their marketing strategies and reach out to younger audiences. In this chapter, a discussion of
the findings with a connection to the theoretical framework will be provided.

6.1 Summary Results

The main topic of this research focused on analyzing Swedish consumers' purchase intention
on TikTok in relation to the factors of trustworthiness, expertise, and parasocial relationships
to see if the factors had a significant effect on purchase intention. To conduct this research the
theory of persuasion and the elaboration likelihood model were used to examine the different
characterizations of influencers based on the influencers’ attributes. Based on past research,
the authors believed there to be a positive relationship between purchase intention and the
believed trustworthiness of the influencer, hence H1 was developed. Further, a believed
relationship between the consumer's perception of the influencer's expertise and the effect on
purchase intention was expected, hence H2 was developed. Lastly, a perceived strong
relationship between the influencer and the consumer was expected to have a significant effect
on the purchase intention, hence H3 was developed to support this claim. In Table 11 the results
are summarized.

40
 Hypothesis Result

H1: Trustworthiness is positively associated with followers’ purchase Supported

intention.

H2: Perceived expertise is positively associated with followers’ purchase Not supported
intention.

H3: Parasocial relationship is positively associated with followers´ Supported

purchase intention.

Table 11. Hypothesis Summary

As presented in Table 10, the majority of the variables possess low significance. One
explanation could be other variables absorbing their effect, however, it is also important to
once again state that Type II errors are a common phenomenon within non-parametric tests.
Thus, a Type II error could be the explanation for the rejection of H2 in Table 11.

6.2 Hypothesis 1
According to the regression model in Table 10, honesty (T1) is the major predictor to purchase
intention. The other two variables within the area of trustworthiness were not proven to be
significant predictors to purchase intention. However, this could possibly originate from how
the questions in the survey were formulated where “I get inspired by TikTokers I believe are
honest” could be perceived as more straightforward than “I get inspired by TikTokers I believe
are credible” and “I get inspired by TikTokers that are acting according to their beliefs and
values” which was connected to the two insignificant variables. Honesty may be a concept
easier to relate to and indicate in a TikTok video, which could explain the variation between
the variables within the area of trustworthiness.

The findings supporting an acceptance of H1 were expected by the authors since multiple prior
studies have indicated that trustworthiness has a major impact on purchase intention as well.
For example, a study by Aggad and Ahmad (2021) stated that trustworthiness is one of the
most influential factors when aiming to create purchase intention originating from influencer
marketing. Since honesty (T1) is proved to be one of the most correlated variables with
purchase intention according to Table 8 and as well as one of the most influential variables for

41
increased purchase intention according to Table 10, creating an honest atmosphere should be a
high priority for influencers performing marketing with the goal to create purchase intention.
Increasing honesty within the marketing can significantly increase purchase intention
according to Table 10. The odds of obtaining high purchase intention when increasing trust is
approximately 2.3 times greater than obtaining one of the lower levels of purchase intention.

As mentioned before, in comparison to the other two questions regarding trustworthiness,

honesty can be perceived as more straightforward and relatable, while the other two concepts
are harder to relate to in the same way. According to Aggad and Ahmad (2021), sincerity,
honesty, or truthfulness are important factors when it comes to trustworthiness. They further
claim that the trustworthiness of influencers had a significant impact on consumers' perceived
credibility, information, purchase intention, e-WOM, and attitudes toward purchase decisions.
Moreover, honesty can therefore be more connected to trustworthiness because the participants
can easier relate to the concept of honesty.

Further connections can be made by looking at Table 4, which shows that 47,83% of the
participants believe that recommended items are more trustworthy than traditional
commercials. A possible connection between the significant variable of honesty (T1) and the
trustworthiness of influencer recommendations can be drawn. Furthermore, a possible parallel
between the participant's value of honesty in relation to purchase intention where almost half
of the participants trust influencer recommendations more than traditional marketing. In turn,
this can be connected to the honesty the participants feel toward the influencer when they are
recommending an item on TikTok, which can be believed to be more trustworthy since the
influencer has proven to be honest in the past.

The positive impact trustworthiness has on purchase intention could be connected to the
statement by Solomon et al. (2019), claiming that influencers possess social power over their
followers and can influence the followers’ consumer behavior. One of the presented factors to
the influencers’ power was information power which referred to the followers believing the
influencer has valuable information, thus acting accordingly to the influencer’s proposal.
Believing someone possesses valuable information could be connected to trustworthiness,
since having trust in another individual’s advice should be required in order to evaluate their
information as important. Thus, based on the acceptance of H1 it could be possible to claim

42
that influencers on TikTok are affecting their followers’ consumer behavior by their
information power.

Ek et al. (2017) claimed that young females in Sweden who emphasize a large trust in the
internet were more likely to have impulse buying behavior. Even though the hypothesis and
following analysis referred to purchase intention and not actual consumption, this could be a
reason why the variable regarding honesty was significant. Since the majority of the
respondents in this research were females (97%) and predominately considered as young (98%
under 30). Hence, it could be possible for these respondents to be more likely to experience
higher levels of purchase intention as well if perceiving trust as an important contributing
factor. However, in order to make any conclusions regarding this, additional research would
be needed.

The findings of H1 offer valuable theoretical contributions to understanding the role of

trustworthiness in consumer behavior, particularly within the context of influencer marketing.
Based on the findings the importance of honesty when aiming to create purchase intention
among consumers on TikTok, is proven to have great significance. Furthermore, this shows the
importance for companies that cooperates with influencers on TikTok, to create honest
marketing with influencers that are recognized for their honesty and trustworthiness online to
enhance the purchase intention among their audience in Sweden. By using the findings of this
study, companies can derive significant benefits by understanding the role of trustworthiness
in Swedish consumer behavior and strategically promoting honesty on TikTok. By prioritizing
honesty in their influencer marketing campaigns, companies can build trust with consumers,
leading to enhanced consumer trust and increased purchase intention.

6.3 Hypothesis 2
The correlation analysis showed a significant association between the variables related to H2,
indicating that there is a relationship between them. However, in the regression analysis, when
controlling for the effects of other variables, the association between the variables related to
H2 may not be significant or may be weaker than previously thought. Thus, hypothesis H2,
which states that expertise is positively associated with purchase intention, is not supported.
The insignificance could occur if there are other variables that are more strongly associated
with the outcome variable and when included in the regression model, overshadow the effects

43
of the variables related to H2. Additionally, the correlation analysis only measures the strength
of the association between variables, while the regression analysis can help determine the
specific contribution of each variable to the outcome variable after controlling for other
variables. It is important to note though, that correlation does not imply causation and there
may be other factors at play that are not accounted for in the analysis. Hence, it may be
necessary to further investigate and analyze the data to fully understand the relationship
between expertise and purchase intention.

There can be several reasons why the literature and the analysis results may differ. One
possibility is that the literature review may have included studies that are based on different
methodologies, samples, contexts, and most probably other social media platforms, leading to
different findings. Since TikTok is relatively new, marketing professionals may not be as active
on this platform as others yet. According to Chu et al. (2022) is the content on TikTok presented
in the format of short videos and Barta et al. (2023) argued that the relationship between TikTok
influencers and their followers is relaxed and personal. This could be a possible explanation as
to why expertise is not a major contributor to purchase intention on TikTok. The followers
might perceive their parasocial relationship to the influencer as more close due to the personal
context and therefore not emphasize expertise as important compared to other platforms where
the context is less personal. The format of short videos could also reduce the importance of
expertise since the influencers do not have as much time for presenting their knowledge as they
do on YouTube for example, which further could reduce expertise’s impact on purchase
intention. If connecting the rejection of H2 to the theory by Solomon et al (2019) regarding
influencers’ social power on their followers’ consumer behavior, expert power could be
perceived as a less influential factor on TikTok for the reasons previously stated.

Expertise might not be a contributing factor to increased purchase intention on TikTok,

however, it could still have an impact on consumers’ purchase behavior. Vrontis et al. (2021)
stated that influencers on social media can affect consumer behavior in different aspects. Even
though consumers might not feel as impacted by influencers with expertise characteristics
regarding their purchase intention, it could still affect their consumer behavior within other
areas. For example, an influencer possessing expertise within an area might not create
immediate purchase intention on TikTok, but it could possibly benefit a brand’s reputation or
create awareness about its existence. Thus, expertise could still be a beneficial marketing tool
for generating future sales.

44
The theoretical contribution of the findings from H2 entails that expertise does not have a
significant impact on purchase intention on TikTok. Even though, H2 was rejected the findings
can still contribute to knowledge of the importance of the other factors that creates purchase
intention on TikTok. The findings show that when using TikTok as a marketing tool, the
variable of expertise is not considered important from the consumers’ point of view, especially
in comparison to parasocial relationships and trustworthiness. As stated above, a personal
relationship with the influencer may seem more important than the expertise the influencer
express in their content. Moreover, the length of the videos can also be a contributing factor to
the small impact expertise had on purchase intention. With this information, companies can
select their marketing channel based on the message they want to deliver or emphasize their
marketing campaigns on the other two factors of this study depending on the aim of the
marketing. By gaining this knowledge from the study companies that want to succeed with
their marketing campaigns on TikTok should focus on applying trustworthiness and parasocial
relationship to be able to create purchase intention amongst their audience in Sweden.

Furthermore, this result is somewhat surprising, as previous research on other platforms has
shown that perceived expertise can positively influence consumer behavior. It is possible that
other factors, such as trustworthiness or parasocial relationship, may have a stronger influence
on purchase intention in this study which generates the low significance of expertise in this
particular study. However, the rejection of H2 could strengthen the importance of addressing
the research gap this study is aiming to cover. Since multiple studies performed on other
platforms have stated expertise as a contributor to increased purchase intention (Li & Peng,
2021; Liu, 2022; Masuda et al., 2022; Lou & Kim, 2019; Wang et al., 2017). The rejection of
H2 within this study aimed for analyzing TikTok specifically, strengthens the argument
regarding the need for analyzing social media platforms individually, since differences among
what is considered as effective influencer marketing strategies are indicated.

45
6.4 Hypothesis 3
According to the correlation model, Parasocial relationships (PR2) have a medium positive
correlation with the dependent variable purchase intention, whilst the four other variables of
parasocial relationship only have been shown to have a low positive correlation (PR1, PR3,
PR4, PR5). The question (PR2), “I get inspired by TikTokers that I believe are natural and
down-to-earth”, also demonstrated significance in the regression model. Moreover, the variable
PR5 proved to be a significant predictor of purchase intention which strengthens the acceptance
of H3 and indicates that parasocial relationship is positively associated with purchase intention.

Connecting to the descriptive statistics in Table 5, where 94,93% indicated that the purpose of
using TikTok is for entertainment purposes. In Table 6, 81,16% of the participants preferred
entertainment as content on TikTok. These statistics can be connected to why PR2 had a
correlation in regard to the other four questions and why only PR2 and PR5 were significant
variables in the regression model. According to Lou and Kim (2019), influencers should rely
on the entertainment value of the content to be able to create parasocial relationships (PSR)
with their followers. Their study discovered the entertainment value of influencer-generated
content and influencer credibility are positively related to PSR. Furthermore, the factors
attractiveness and similarity showed a strong relation of the formation of PSR between
influencer and follower, where PSR was positively associated with purchase intention in their
study. Based on the findings, parallels can be drawn from past studies to the results given in
this study where both attractiveness and “down-to-earth” had significance in Table 10.
Question PR2 could then be connected to the term “similarity” which Lou and Kim use in their
study. The formulation of the other questions in their study under parasocial relationships
appears to be more connected to a believed friendship and not feelings of similarities, thus can
be connected to acting natural and “down-to-earth”.

Additionally, a conclusion can be drawn between the high percentages that indicated the
importance of the entertainment factor and the positive relation between purchase intention and
parasocial relationships. Further, both attractiveness and “down-to-earth” plays an important
role in creating PSR between an influencer and the followers which in turn has proven to have
a significant impact on purchase intention. Therefore, it can be stated that by focusing on PSR,
especially the factors of attractiveness and “down-to-earth”, the influencer will generate higher
purchase intention with their content on TikTok.

46
The positive impact of parasocial relationships on purchase intention can be connected to
Solomon et al. (2019) stating that influencers can possess social power over their followers
through referral power. Referral power refers to when the followers mirror the influencer’s
behavior due to their admiration of the influencer, which possible could be connected to
experiencing a parasocial relationship with the influencer. The perception of having a close
relationship with an influencer, as if it was a real friend, probably indicates a strong liking of
the influencer as well. Thus, it is not impossible that admiration of the influencer also exists
due to their likable characteristics, which would indicate the possession of referral power by
the influencer.

The theoretical contribution of the findings from H3 supports that parasocial relationships are
positively associated with purchase intention. This suggests that companies can leverage the
power of parasocial relationships by partnering with celebrities or influencers to endorse their
products since they have the power to create a believed friendship and closeness to their
audience. Moreover, by developing content with the basis of being entertaining and including
the factors of attractiveness and the influencer's possibility to express themself as natural and
“down-to-earth”, the possibility of endorsing products that create purchase intention is
significantly higher. By strategically leveraging parasocial relationships, companies can foster
stronger connections with their target audience and increase the effectiveness of their
marketing efforts on TikTok. However, it is important to note that these findings only show a
correlation and do not necessarily imply causation. Further research is needed to determine the
directionality and underlying mechanisms of this relationship. Nonetheless, the findings
provide valuable insights into the influence of parasocial relationships on Swedish consumer
behavior.

47
7. Conclusion, Implication, and Limitations

7.1 Conclusion
This study aimed to investigate Swedish consumers' purchase intention on TikTok in relation
to the factors of trustworthiness, expertise, and parasocial relationships to see if the factors had
a significant effect on purchase intention. Based on the analysis conducted, it can be concluded
that trustworthiness and parasocial relationships are positively associated with purchase
intention, while expertise does not have a significant impact on purchase intention. The
correlation analysis showed a significant positive relationship between expertise and purchase
intention, however, this was not supported in the regression analysis. It is important to note that
the results are based on the specific data and methodology used in the analysis, and may not be
generalizable to other populations or contexts.

The conclusion of the research suggests that when a company intends to promote the brand
using influencers on TikTok to target Swedish consumers, it should consider incorporating the
elements of trustworthiness and parasocial relationships. The findings from this research
suggest that by emphasizing honesty in communication or collaborating with influencers
known for their honesty, the purchase intention can be significantly higher. Moreover, the
research establishes that the perceived parasocial relationship also has a significant impact on
purchase intention. This highlights the importance for both companies and influencers to
consider this factor when developing their marketing strategies. In line with the theoretical
framework, employing techniques such as positioning the camera at eye level, maintaining eye
contact through the camera, engaging in private and informal conversations, and greeting the
audience, can evoke a sense of friendship and trust. These findings suggest that such strategies
can lead to an increased purchase intention on the app.

Perceived expertise was proven to have no significance on purchase intention on the platform
TikTok. Firstly, the low significance can be explained by the stronger associations
demonstrated by other variables with respect to purchase intention. This circumstance may
have led to an overshadowing effect, where the variables related to perceived expertise were
reduced in their impact. Secondly, followers might perceive their parasocial relationship with
the influencer as more important due to the personal context of TikTok and therefore, not

48
emphasize expertise as valuable. However, previous studies have proven expertise to have a
significant effect on purchase intention, which could be explained by the cultural differences
among the platforms, where TikTok is more focused on short-from entertaining videos. The
format of short videos could reduce the importance of expertise since the influencers do not
have as much time for presenting their knowledge as they do on, for example, YouTube, which
further could reduce expertise’s impact on purchase intention.

Overall, this study contributes to the understanding of the market of influencer marketing on
TikTok and provides insights into the factors that influence Swedish consumers' purchase
intention. The findings of this study can be useful for marketing professionals and influencers
to develop effective influencer marketing strategies on TikTok.

7.2 Implications

7.2.1 Theoretical Implications

There is an overall lack of research on the platform TikTok and more specifically, research
regarding marketing on the platform. There is a wide range of research regarding social media
and marketing on other social media platforms but there is insufficient literature addressing the
platform TikTok. Hence, this study aims to provide a better understanding of the influence of
TikTok as a social media platform on consumers’ purchase intention, filling the existing gap
in research on this topic. In the extent of theoretical implication in this study, the authors have
applied an adapted version of the theory of persuasion by Masuda et al. (2022) to strengthen
the understanding of the relationship between the influencer characteristics of trustworthiness,
expertise, and parasocial relationship and the purchase intention of Swedish consumers.

7.2.2 Practical Implications

The results suggest that trustworthiness is positively associated with purchase intention. This
implies that influencers should focus on building trust with their followers to increase purchase
intention. The results also suggest that parasocial relationships are positively associated with
purchase intention. Thus, companies should focus on collaborating with influencers possessing
strong trustworthiness and parasocial relationships with their followers to increase purchase
intention. However, the results did not find a significant positive association between expertise

49
and purchase intention. This implies that influencers should not solely rely on their expertise
to increase purchase intention on TikTok.

7.3 Limitations and Future Research

As for other studies, this has limitations as well. Quantitative research includes limitations for
instance potential constraints, weaknesses, or biases. In addition, a quantitative research that
uses a survey method does not allow for further elaboration on the given answer in contrast to
a qualitative study. Hence, for future research, a qualitative research could be performed for a
more in depth analysis of the consumer perspective. For example, why expertise is perceived
as less influential on TikTok as found in this study. A large limitation that opens opportunities
for future research is the limitation of the sample size. In this study, the researcher has only
examined Swedish consumers which could cause the external validity of the results to be
influenced to some extent. Additionally, due to the Swedish legislation (1949:381), only
consumers aged 18 or older have been examined. The population of this study has been reached
mainly from the authors' social media platforms which may indicate a risk that the sample size
differs from the age distribution on TikTok. Moreover, the sample size is limited in regards to
the population and mainly consists of respondents aged between 23-26.

For future research, a more diverse sample could be applied to compare how different age
groups evaluate trustworthiness, expertise and parasocial relationships. In addition, since
adolescences are a major group on TikTok, this age group could be analyzed individually to
gain an understanding of how influencer marketing is affecting their consumer behavior. This
research considers the factors of trustworthiness, expertise, and parasocial relationship. For
additional future research, other factors can be investigated, for example, other influencer
characterizations. Furthermore, entertainment was indicated to be the most preferred content
on TikTok however, for future research a study could be conducted to investigate which types
of content are creating the most purchase intention. Moreover, since this study only examined
Swedish consumers, results may differ across different cultures and countries as influencer
marketing strategies may vary depending on cultural and societal norms.

50
8. References
Aggad, K. K., & Ahmad, F. S. (2021). Investigates the Impact of Social Media Influencers’
Personality, Content, and Trustworthiness on Consumers’ Purchase Intention and eWOM.
International Journal of Academic Research in Business and Social Sciences, 11(12).
https://doi.org/10.6007/IJARBSS/v11-i12/11782

Alin, A. (2010). Multicollinearity. Wiley Interdisciplinary Reviews. Computational Statistics,

2(3), 370–374. https://doi.org/10.1002/wics.84

Anderson, D. R., Sweeney, D. J., Williams, T. A., Freeman, J., Shoesmith, E. (2014).
Statistics for business & economics. Cengage Learning.

Bacev-Giles, C., & Haji, R. (2017). Online first impressions: Person perception in social
media profiles. Computers in Human Behavior, 75, 50–57.
https://doi.org/10.1016/j.chb.2017.04.056

Bao, T. T., & Chang, T. L. S. (2016). The product and timing effects of eWOM in viral
marketing. International Journal of Business, 21(2), 99.
https://ijb.cyut.edu.tw/var/file/10/1010/img/862/V212-1.pdf

Barta, S., Belanche, D., Fernández, A., & Flavián, M. (2023). Influencer marketing on
TikTok: The effectiveness of humor and followers’ hedonic experience. Journal of Retailing
and Consumer Services, 70, 103149–. https://doi.org/10.1016/j.jretconser.2022.103149

Bell, E., & Bryman, A. (2007). The ethics of management research: an exploratory content
analysis. British journal of management, 18(1), 63-77.

Bhattacharya, S., Gaurav, K., & Ghosh, S. (2019). Viral marketing on social networks: An
epidemiological perspective. Physica A, 525, 478–490.
https://doi.org/10.1016/j.physa.2019.03.008

Bi, N. C., & Zhang, R. (2022). “I will buy what my ‘friend’ recommends”: the effects of
parasocial relationships, influencer credibility and self-esteem on purchase intentions.

51
Journal of Research in Interactive Marketing, 17(2), 157–175. https://doi.org/10.1108/JRIM-
08-2021-0214

Buchholz, K. (Oct 7, 2022). The Rapid Rise of TikTok. Statista. Retrieved April 24th 2023
from: https://www.statista.com/chart/28412/social-media-users-by-network-amo/

Chu, S.-C., Deng, T., & Mundel, J. (2022). The impact of personalization on viral behavior
intentions on TikTok: The role of perceived creativity, authenticity, and need for uniqueness.
Journal of Marketing Communications, 1–20.
https://doi.org/10.1080/13527266.2022.2098364

Chung, S., & Cho, H. (2017). Fostering Parasocial Relationships with Celebrities on Social
Media: Implications for Celebrity Endorsement. Psychology & Marketing, 34(4), 481–495.
https://doi.org/10.1002/mar.21001

Collis, J., & Hussey, R. (2014). Business Research. Palgrave Macmillan Higher
Education.

Dencheva, V. (2023a, January 6). TikTok marketing - statistics & facts. Statista.
https://www.statista.com/topics/8309/tiktok-marketing/#topicOverview

Dencheva, V. (2023b, January 6). Share of selected platforms in influencer marketing

spending in the United States from 2019 to 2024. Statista. https://www-statista-
com.proxy.library.ju.se/statistics/1329237/influencer-marketing-spending-platform-usa/

Edgar, T. W., & Manz, D. O. (2017). Exploratory Study. Research Methods for Cyber
Security, 95-130. https://doi.org/10.1016/B978-0-12-805349-2.00004-2

Ek, S., Foster, T., & Wallström, Å. (2017). Impulse buying tendencies among online
shoppers in Sweden. Journal of Research in Interactive Marketing, 11(4), 416–431.
https://doi.org/10.1108/JRIM-05-2016-0054

52
Emerson, R. W. (2015). Convenience Sampling, Random Sampling, and Snowball Sampling:
How Does Sampling Affect the Validity of Research? Journal of Visual Impairment &
Blindness, 109(2), 164–168. https://doi.org/10.1177/0145482X1510900215

Erkan, I., & Evans, C. (2018). Social media or shopping websites? The influence of eWOM
on consumers’ online purchase intentions. Journal of Marketing Communications, 24(6),
617–632. https://doi.org/10.1080/13527266.2016.1184706

Foxall, G. R. (2015). Consumer behaviour : a practical guide. Routledge.

https://doi.org/10.4324/9781315736082

Geyser, W. (2023, 24 March). What is an Influencer? - Social Media Influencers Defined.

Influencer Marketing Hub. https://influencermarketinghub.com/what-is-an-influencer/#toc-0

Gomes, M. A., Marques, S., & Dias, Á. (2022). The impact of digital influencers’
characteristics on purchase intention of fashion products. Journal of Global Fashion
Marketing, 13(3), 187–204. https://doi.org/10.1080/20932685.2022.2039263

Gutierrez., R. G. (2010). Stata. Wiley Interdisciplinary Reviews. Computational Statistics,

2(6), 728–733. https://doi.org/10.1002/wics.116

Haenlein, M., Anadol, E., Farnsworth, T., Hugo, H., Hunichen, J., & Welte, D. (2020).
Navigating the New Era of Influencer Marketing: How to be Successful on Instagram,
TikTok, & Co. California Management Review, 63(1), 5–25.
https://doi.org/10.1177/0008125620958166

Haglund, A. (2023, March 10). Regeringen duckar om förbud mot Tiktok. SvD.
https://www.svd.se/a/2Bo9yx/regeringen-duckar-om-forbud-mot-tiktok

Harpe, S. E. (2015). How to analyze Likert and other rating scale data. Currents in Pharmacy
Teaching and Learning, 7(6), 836–850. https://doi.org/10.1016/j.cptl.2015.08.001

Harrigan, P., Soutar, G. N., Weismueller, J., & Wang, S. (2020). Influencer endorsements:
How advertising disclosure and source credibility affect consumer purchase intention on

53
social media. Australasian Marketing Journal, 28(4), 160–170.
https://doi.org/10.1016/j.ausmj.2020.03.002

Harrison, Kemp, R., Brace, N., & Snelgar, R. (2021). SPSS for psychologists (Seventh
edition). Macmillan International Higher Education.

Hilbe, J. (2009). Logistic regression models. Chapman & Hall/CRC.

https://doi.org/10.1201/9781420075779

Hjort, K., & Lantz, B. (2016). The impact of returns policies on profitability: A fashion e-
commerce case. Journal of Business Research, 69(11), 4980–4985.
https://doi.org/10.1016/j.jbusres.2016.04.064

Hovland, C., Janis, I., & Kelley, H. (1953). Communication and Persuasion: Psychological
Studies of Opinion Change. Oxford University Press.

Hwang, K., & Zhang, Q. (2018). Influence of parasocial relationship between digital
celebrities and their followers on followers’ purchase and electronic word-of-mouth
intentions, and persuasion knowledge. Computers in Human Behavior, 87, 155–173.
https://doi.org/10.1016/j.chb.2018.05.029

Ismagilova, E., Dwivedi, Y. K., Slade, E., & Williams, M. D. (2017). Electronic Word of
Mouth (eWOM) in the Marketing Context A State of the Art Analysis and Future Directions
(1st ed. 2017.). Springer International Publishing. https://doi.org/10.1007/978-3-319-52459-7

Kanaveedu, A., & Kalapurackal, J. J. (2022). Influencer Marketing and Consumer behavior:
A Systematic Literature Review. Vision (New Delhi, India).
https://doi.org/10.1177/09722629221114607

Koay, K. Y., Cheung, M. L., Soh, P. C-H., & Teoh, C. W. (2022). Social media influencer
marketing: the moderating role of materialism. European Business Review, 34(2), 224–243.
https://doi.org/10.1108/EBR-02-2021-0032

54
Koetsier, J. (2021). Top 10 Most Downloaded Apps And Games Of 2021: TikTok, Telegram
Big Winners. Forbes. Retrieved March 27 from:
https://www.forbes.com/sites/johnkoetsier/2021/12/27/top-10-most-downloaded-apps-and-
games-of-2021-tiktok-telegram-big-winners/?sh=31075a2f3a1f

Laksamana, P. (2018). Impact of Social Media Marketing on Purchase Intention and Brand
Loyalty: Evidence from Indonesia’s Banking Industry. International Review of Management
and Marketing, 8(1).

Leite, F. P., & Baptista, P. P. (2022). The effects of social media influencers’ self-disclosure
on behavioral intentions: The role of source credibility, parasocial relationships, and brand
trust. Journal of Marketing Theory and Practice, 30(3), 295–311.
https://doi.org/10.1080/10696679.2021.1935275

Lenkert, E. (2023). What is a content creator and how to become one?. Adobe.com.
https://www.adobe.com/express/learn/blog/content-creator#what-is-a-content-creator

Li, Y., & Peng, Y. (2021). Influencer marketing: purchase intention and its antecedents.
Marketing Intelligence & Planning, 39(7), 960–978. https://doi.org/10.1108/MIP-04-2021-
0104

Liu, M. (2022). Determining the Role of Influencers’ Marketing Initiatives on Fast Fashion
Industry Sustainability: The Mediating Role of Purchase Intention. Frontiers in Psychology,
13, 940649–940649. https://doi.org/10.3389/fpsyg.2022.940649

Lou, C., & Kim, H. K. (2019). Fancying the New Rich and Famous? Explicating the Roles of
Influencer Content, Credibility, and Parental Mediation in Adolescents’ Parasocial
Relationship, Materialism, and Purchase Intentions. Frontiers in Psychology, 10, 2567–2567.
https://doi.org/10.3389/fpsyg.2019.02567

Magano, J., Au-Yong-Oliveira, M., Walter, C. E., & Leite, Â. (2022). Attitudes toward
Fashion Influencers as a Mediator of Purchase Intention. Information (Basel), 13(6), 297–.
https://doi.org/10.3390/info13060297

55
Maheshwari, S., & Holpuch, A. (2023, April 4). Why Countries Are Trying to Ban TikTok.
The New York Times. https://www.nytimes.com/article/tiktok-ban.html

Masuda, H., Han, S. H., & Lee, J. (2022). Impacts of influencer attributes on purchase
intentions in social media influencer marketing: Mediating roles of characterizations.
Technological Forecasting & Social Change, 174, 121246–.
https://doi.org/10.1016/j.techfore.2021.121246

McNulty, K. (2021). Handbook of regression modeling in people analytics : with examples in

R and Python. CRC Press.

Mishra, A. Maheswarappa, S. S., Maity, M., & Samu, S. (2018). Adolescent’s eWOM
intentions: An investigation into the roles of peers, the Internet and gender. Journal of
Business Research, 86, 394–405. https://doi.org/10.1016/j.jbusres.2017.04.005

Morris, S. (2021, April 9). 20 Products that Have Sold Out Because of TikTok Popularity.
Newsweek. https://www.newsweek.com/20-products-sold-out-because-tiktok-popularity-
1581946

Morwitz, V. (2014). Consumers' Purchase Intentions and their Behavior. Foundations and
Trends® in Marketing, 7(3), 181-230. http://dx.doi.org/10.1561/1700000036

Murray, J. (2013). Likert Data: What to Use, Parametric or Non-Parametric?. International

Journal of Business and Social Science,

Nemoto, T., & Beglar, D. (2014). Likert-scale questionnaires. In JALT 2013 conference
proceedings (pp. 1-8). International Journal of Business and Social Science, 4(11).

O’Connell, A. A. (2006). Logistic regression models for ordinal response variables. SAGE.

Ohanian, R. (1990). Construction and Validation of a Scale to Measure Celebrity Endorsers’

Perceived Expertise, Trustworthiness, and Attractiveness. Journal of Advertising, 19(3), 39–
52. https://doi.org/10.1080/00913367.1990.10673191

56
Onu, C. A., Nwaulune, J., Adegbola, E. A., & Nnorom, G. (2019). The effect of celebrity
physical attractiveness and trustworthiness on consumer purchase intentions: A study on
Nigerian consumers. Management Science Letters, 9(12), 1965–1976.
https://doi.org/10.5267/j.msl.2019.7.009

Osborne, J. W. (2015). Best practices in logistic regression. SAGE.

Osborne‐Gowey, J. (2014). What is Social Media. Fisheries, 39(2), 55–55.

https://doi.org/10.1080/03632415.2014.876883

Petty, R. E., & Cacioppo, J. T. (1981). Attitudes and persuasion: Classic and Contemporary
Approaches. William C Brown Company Publishers.

Petrescu, M. (2014). Viral marketing and social networks (First edition.). Business Expert
Press.

Reinard, J. C. (2006). Communication research statistics. SAGE.

Reinikainen, H., Munnukka, J., Maity, D., & Luoma-aho, V. (2020). “You really are a great
big sister” - parasocial relationships, credibility, and the moderating role of audience
comments in influencer marketing. Journal of Marketing Management, 36(3-4), 279–298.
https://doi.org/10.1080/0267257X.2019.1708781

Ryan, C. (2021). Non-parametric Tests. In Data Science with R for Psychologists and
Healthcare Professionals. Taylor & Francis Group.

Sammis, K., Lincoln, C., Pomponi, S., Ng, J., Gassmann Rodriguez, E., & Zhou, J. (2016).
Influencer marketing for dummies (1st edition). Wiley.

Schouten, A., Janssen, L., & Verspaget, M. (2020). Celebrity vs. Influencer endorsements in
advertising: the role of identification, credibility, and Product-Endorser fit. International
Journal of Advertising, 39(2), 258–281. https://doi.org/10.1080/02650487.2019.1634898

57
Scrivener, E. (2022, March 30). Lookfantastic launches a beauty box with five of the most
viral beauty products of the past year. Daily Mail.
https://www.dailymail.co.uk/femail/article-10663807/Lookfantastic-launches-beauty-box-
five-viral-beauty-products-past-year.html

Shao, J.-H. & Jing, R.-Z. (2022). Viral marketing strategies with dual incentives. Electronic
Commerce Research and Applications, 54, 101180–.
https://doi.org/10.1016/j.elerap.2022.101180

Sokolova, K., & Kefi, H. (2020). Instagram and YouTube bloggers promote it, why should I
buy? How credibility and parasocial interaction influence purchase intentions. Journal of
Retailing and Consumer Services, 53, 101742–.
https://doi.org/10.1016/j.jretconser.2019.01.011

Solomon, M. R., Askegaard, S., Hogg, M. K., & Bamossy, G. J. (2019). Consumer behavior :
a European perspective (7th edition.). Pearson.

Spears, N., & Singh, S. N. (2004). Measuring Attitude toward the Brand and Purchase
Intentions. Journal of Current Issues and Research in Advertising, 26(2), 53–66.
https://doi.org/10.1080/10641734.2004.10505164

Statista Research Department. (2023, January 9). Influencer marketing spending in Sweden
from 2016 to 2022. Statista. https://www-statista-
com.proxy.library.ju.se/statistics/1356217/influencer-marketing-spending-sweden/

Svenskarna och internet. (2022). Svenskarna och internet 2022.

https://svenskarnaochinternet.se/rapporter/svenskarna-och-internet-2022/sociala-
medier/#tiktok

Sweeney, J. C., Soutar, G. N., & Mazzarol, T. (2008). Factors influencing word of mouth
effectiveness: receiver perspectives. European Journal of Marketing, 42(3/4), 344–364.
https://doi.org/10.1108/03090560810852977

58
TikTok. (2022). Camera Tools. TikTok. Retrieved March 27 from:
https://support.tiktok.com/en/using-tiktok/creating-videos/camera-tools

Tuten, T. L. (2021). Social media marketing (4th edition). SAGE Publishing.

Vittinghoff, E., & McCulloch, C. E. (2007). Relaxing the Rule of Ten Events per Variable in
Logistic and Cox Regression. American Journal of Epidemiology, 165(6), 710–718.
https://doi.org/10.1093/aje/kwk052

Vrontis, D., Makrides, A., Christofi, M., & Thrassou, A. (2021). Social media influencer
marketing: A systematic review, integrative framework and future research agenda.
International Journal of Consumer Studies, 45(4), 617–644.
https://doi.org/10.1111/ijcs.12647

Wadgave, U., & Khairnar, M. R. (2016). Parametric Tests for Likert Scale: For and Against.
Asian Journal of Psychiatry, 24, 67–68. https://doi.org/10.1016/j.ajp.2016.08.016

Wang, S., Kao, G. H.-Y., & Ngamsiriudom, W. (2017). Consumers’ attitude of endorser
credibility, brand and intention with respect to celebrity endorsement of the airline sector.
Journal of Air Transport Management, 60, 10–17.
https://doi.org/10.1016/j.jairtraman.2016.12.007

Wiley, D. (January 11, 2022).6 Red-Hot Influencer Marketing Predictions For 2022. Forbes.
https://www.forbes.com/sites/forbesagencycouncil/2022/01/11/6-red-hot-influencer-
marketing-predictions-for-2022/?sh=6b16c7a41f28

Wilkie, C., & Feiner, L. (2023, March 1). New TikTok ban bill passes key House committee
on a party-line vote. CNBC. https://www.cnbc.com/2023/03/01/new-tiktok-ban-bill-passes-
key-house-committee-on-a-party-line-vote.html

Ye, G., Hudders, L., De Jans, S., & De Veirman, M. (2021). The value of influencer
marketing for business: A bibliometric analysis and managerial implications. Journal of
Advertising, 50(2), 160-178.

59
Yesiloglu, S. & Costello, J. (2021). Influencer marketing : building brand communities and
engagement (Yesiloglu & J. Costello, Eds.). Routledge.

Zhou, S., McCormick, H., Blazquez, M., & Barnes, L. (2019). eWOM: The Rise of the
Opinion Leaders. In Social Commerce (pp. 189–212). Springer International Publishing.
https://doi.org/10.1007/978-3-030-03617-1_11

SFS 1949:381. Om underårigs omyndighet. https://www.riksdagen.se/sv/dokument-

lagar/dokument/svensk-forfattningssamling/foraldrabalk-1949381_sfs-1949-381#K9

60
Appendix

Appendix 1: The Rapid Rise of TikTok

Appendix 2: Ordered Logistics Regression Odds Ratio

61
Appendix 3: Mean Age of Respondents

Appendix 4: Skewness and Kurtosis Tests for Normality

62
Appendix 5: Survey Introduction Text
Hello!

This survey is aimed at those who have TikTok.

The survey consists of 23 questions and takes about 2 minutes to complete.

We are three students in the final semester of the Marketing Management program at Jönköping
International Business School. We are working on our thesis on how product recommendations
on the TikTok platform affect consumer purchase intentions. The purpose of this survey is to
collect data on the trustworthiness, expertise, and parasocial relationships of a TikToker and
how it affects purchase intention, in order to later contribute to the knowledge of success factors
in influencer marketing on TikTok.

Participation is completely voluntary and anonymous. By participating, you give us permission

to use your answers in our study. Your personal information will only be used for scientific
purposes in our thesis and will be treated in accordance with GDPR regulations. No data will
be saved after the publication of our thesis.

The thesis writers are: Ellen Johansson, Filippa Nilsson & Johanna Nilsson
Supervisor for this thesis is: Fatih Cura

For other questions and comments, please contact us.

Thank you in advance!

63
Appendix 6: Survey Items

Item Source

Q1: Gender
○ Man
○ Woman
○ Other

Q2: Age
○ (write a number)

Q3: Education
○ Elementary school
○ High School
○ University
○ Other education after high school

Q4: Income
○ Less than 12,500 SEK
○ 12,500 - 20,000 SEK
○ 20,001-30,000 SEK
○ 30,001 - 40,000 SEK
○ More than 40,000 SEK

Q5: How often do you use TikTok?

○ Multiple times a day
○ Once a day
○ Multiple times a week
○ Few times a week
○ Less often

TikTok usage and Purchase Intention:

Q6: I use TikTok for;

(Possible to choose more than one alternative)
a. Inspiration
b. Entertainment
c. Create videos

64
 d. Product recommendations
e. Gaining knowledge
f. Other

Q7: I prefer the following types of content/content on TikTok:

(Possible to choose more than one alternative)
○ Entertainment
○ Dancing
○ Pranks
○ Health and workout tips
○ DIY / rebuilding
○ Hair and Makeup
○ Fashion
○ Recipes
○ Lifehacks
○ Animals
○ Travels
○ Other

Q8: I get inspired by the content I see on TikTok:

1-5 (1=Disagree completely - 5= Agree completely)

Q9: I feel that I want to purchase products/services I see on TikTok: *Dependent variable*
(this does not necessarily mean that you actually purchased the product/service,
more if you have felt that you wanted to)
1-5 (1=Disagree completely - 5= Agree completely)

Q10: I purchase product/services I have seen on TikTok:

1-5 (1=Disagree completely - 5= Agree completely)

Q11: I feel that the recommendations I see on TikTok (not sponsored posts)
are more trustworthy than traditional commercials:
1-5 (1=Disagree completely - 5= Agree completely)

Trustworthiness:

Q12: I get inspired by TikTokers I believe are honest: Masuda et al. (2022)
1-5 (1=Disagree completely - 5= Agree completely)

Q13: I get inspired by TikTokers I believe are reliable: Masuda et al. (2022)
1-5 (1=Disagree completely - 5= Agree completely)

65
Q14: I get inspired by TikTokers that are acting according to their beliefs and Masuda et al. (2022)
values:
1-5 (1=Disagree completely - 5= Agree completely)

Expertise:

Q15: I get inspired by TikTokers that are proficient: Masuda et al. (2022)
1-5 (1=Disagree completely - 5= Agree completely)

Q16: I get inspired by TikTokers that I believe are knowledgeable in the area Masuda et al. (2022)
of what they are recommending:
1-5 (1=Disagree completely - 5= Agree completely)

Q17: I get inspired by TikTokers I believe are experts in a special area: Masuda et al. (2022)
1-5 (1=Disagree completely - 5= Agree completely)

Parasocial Relationships:

Q18: I get inspired by TikTokers which makes me feel comfortable as if it was Masuda et al. (2022)
a close friend:
1-5 (1=Disagree completely - 5= Agree completely)

Q19: I get inspired by TikTokers that I believe are natural and down-to-earth: Masuda et al. (2022)
1-5 (1=Disagree completely - 5= Agree completely)

Q20: I get inspired by TikTokers that understand more about areas that I want Masuda et al. (2022)
to learn more about:
1-5 (1=Disagree completely - 5= Agree completely)

Q21: I get inspired by TikTokers whom I would like to meet in real life: Masuda et al. (2022)
1-5 (1=Disagree completely - 5= Agree completely)

Q22: I get inspired by TikTokers that I find attractive: Masuda et al. (2022)
○ 1-5 (1=Disagree completely - 5= Agree completely)

Q23: What was this survey about? (control question)

○ Mathematics in elementary school
○ Sweden's food-culture
○ Influencers on TikTok
○ Instagram's algorithm

66
Appendix 7: Variable Names

Area of research Variable Question

Demographics Gender 1. Gender:

Age 2. Age:

Education 3. Education:

Income 4. Income:

Usage_freq 5. How often do you use TikTok?

TikTok usage and Purpose_alternative (dummy 6. I use TikTok for:

purchase intention variable for each alternative)

Preffered_alternative 7. I prefer the following types of

(dummy variable for each content/contents on TikTok:
alternative)

Inspiration 8. I get inspired by the content I

see on TikTok:

Purchase_intention 9. I feel that I want to purchase

products/services I see on
TikTok:

Purchase 10. I purchase product/services i

have seen on TikTok:

Recommendations 11. I feel that the recommendations

I see on TikTok (not sponsored
posts) are more trustworthy than
traditional commercials:

Trustworthiness T1 12. I get inspired by TikTokers I

believe are honest:

67
 T2 13. I get inspired by TikTokers I
believe are reliable:

T3 14. I get inspired by TikTokers that

are acting according to their
beliefs and values:

Expertise E1 15. I get inspired by TikTokers that

are proficient:

E2 16. I get inspired by TikTokers that

I believe are knowledgeable in
the area of what they are
recommending:

E3 17. I get inspired by TikTokers I

believe are experts in a special
area:

Parasocial PR1 18. I get inspired by TikTokers

relationship which makes me feel
comfortable as if it was a close
friend:

PR2 19. I get inspired by TikTokers that

I believe are natural and down-
to-earth:

PR3 20. I get inspired by TikTokers that

understand more about areas
that I want to learn more about:

PR4 21. I get inspired by TikTokers

whom i would like to meet in
real life:

68
PR5 22. I get inspired by TikTokers that
I find attractive:

Control 23. What was this survey about?

(control question)

69
Appendix 8: Frequency and Percentage of Survey Answers

70
71