The current issue and full text archive of this journal is available on Emerald Insight at:

https://www.emerald.com/insight/2040-7122.htm

The impact of Internet celebrity Internet

celebrity
characteristics on followers’ characteristics

impulse purchase behavior:

the mediation of attachment and 483
parasocial interaction Received 4 September 2020
Revised 11 March 2021
6 May 2021
Tser Yieth Chen 20 May 2021
Accepted 21 May 2021
Graduate Institute of International Business, National Taipei University,
New Taipei City, Taiwan
Tsai Lien Yeh
Department of International Business, Ming Chuan University, Taipei, Taiwan, and
Fang Yu Lee
Graduate Institute of International Business, National Taipei University,
New Taipei City, Taiwan

Abstract
Purpose – This study aims to investigate the relationship between Internet celebrity characteristics and the
follower’s impulse purchase behavior in YouTuber. Attachment and parasocial interaction are mediating
variables concerning the impact of Internet celebrity characteristics on followers’ impulse purchase behavior.
Design/methodology/approach – This study conducted an online survey of YouTube users in Taiwan and
utilized quota sampling with 500 samples to examine the influence of two types of Internet celebrities on
impulse purchase behavior.
Findings – As to the empirical results, the main path indicated that the self-disclosure Internet celebrity
positively affected the attachment, which mainly positively affected the impulse purchase behavior. The
second path showed that the expertise-knowledge Internet celebrity positively affected the attachment, which
positively affected the impulse purchase behavior.
Practical implications – YouTube marketers should proceed prudently with the market segmentation and
choose the appropriate type of Internet celebrities who are suitable for the product image to differentiate
marketing. Empirical results can aid marketers in selecting a product-endorser, and enhance consumers’
purchasing effect on product advertisements in interactive marketing.
Originality/value – The novelty of this study is to explore the mediation effect of the impact of Internet
celebrity characteristics on followers’ impulse purchase behavior in interactive marketing. The explaining
mechanism of attachment and parasocial interaction is promised to be highlighted as the contribution of this
study to the extant literature. This study constructs a theoretical mechanism between attachment theory and
parasocial interaction theory and then can be used as a theoretical lens for designing successful social media
strategies and explaining social media brand relationships.
Keywords Internet celebrity, Self-disclosure, Expertise-knowledge, Attachment, Parasocial interaction,
Impulse purchase behavior
Paper type Research paper

Funding details: There is no any funding source in this study. Journal of Research in Interactive
Ethical approval: This article does not contain any studies with human participants or animals Marketing
Vol. 15 No. 3, 2021
performed by any of the authors. pp. 483-501
Conflict of interest disclosure statement: As for conflict of interest, the authors declare no conflict of © Emerald Publishing Limited
2040-7122
interest. DOI 10.1108/JRIM-09-2020-0183
JRIM Introduction
15,3 Internet celebrities have been shown to affect the decisions of their social media followers
(Akdevelioglu and Kara, 2020; Al-Emadi and Ben Yahia, 2020). Therefore, investigating
the role of Internet celebrities in impulse purchase behavior is of great interest to Internet
marketing managers. Growing evidence shows that followers of these celebrities on
Internet platforms are prone to impulse buying behavior owing to parasocial relationships
(Graham and Wilder, 2020; Styv
en et al., 2017; Xiang et al., 2016). Recent literature indicated
484 that Internet celebrities attract various followers through individual characteristics via the
cognitive processing mechanism (Akdevelioglu and Kara, 2020; Ferchaud et al., 2018; Vander
Schee et al., 2020). Cognitive processing mechanism is when the user assesses the “scent”
provided by the link’s text or icon, determining whether the link will lead to desired content
(Ong et al., 2017). However, existing studies on online impulse buying were limited when
illustrating the behavior of followers in parasocial relationships as social relationship factors
(Chung and Cho, 2017; Cole and Leets, 1999; Gong and Li, 2017; Rosaen and Dibble, 2016;
Stever, 2013; Xiang et al., 2016). New explanatory mechanisms are needed to fill the research
gap. In this study, attachment and parasocial interaction are introduced as mediation effects
to examine the influence of parasocial relationship factors on the formation of impulse buying
behavior.
In particular, parasocial relationships, as a kind of psychological relationship experienced
by an audience in mediated encounters with performers, definitely play an important role in
interactive marketing (Labrecque, 2014). The interactive role of parasocial relationships,
including attachment and parasocial interaction, should be highlighted in the framework of
the impact of Internet celebrity characteristics on followers’ impulse purchase behavior from
the interactive marketing perspective (Qin, 2020; Xiang et al., 2016). The mediation effect
explains the mechanism underlying the proposed relationship antecedents, and outcomes are
important for recent knowledge advancement and/or managerial implications. However,
recent literature did not address the mediating effect of the impact of Internet celebrity
characteristics on impulse purchase behavior, which formed a research gap (Chung and Cho,
2017; Corr^ea et al., 2020; Ferchaud et al., 2018; Gong and Li, 2017). The explanatory
mechanisms of attachment and parasocial interaction are the contribution of this study to
extant literature.
Attachment (AT) means that followers believe that an idol will meet their affective demand
as fans and then develop an imaginary relationship with them (Yue, 2010). Individuals’
attachment experiences influence their thoughts, their actions and interactions with others
(Hinson et al., 2019; Moulard et al., 2015). Therefore, consumers’ attachment to an Internet
celebrity can be expressed in continual emotional bonding with one brand, and by identifying
with the targeted Internet celebrity (Corr^ea et al., 2020; Utz, 2015). Followers’ sense of social
closeness with the targeted Internet celebrity in their life can be assessed based on attachment
theory (Cole and Leets, 1999; Hinson et al., 2019). Followers with similar characteristics may be
more likely to feel attachment to a targeted Internet celebrity than dissimilar ones, which is
known as homophily (Kim and Altmann, 2017; Rosaen and Dibble, 2016).
A parasocial interaction (PSI) is a type of interaction in a face-to-face personal relationship
between an Internet celebrity and his or her followers (Xiang et al., 2016). Parasocial
interaction highlighted the “illusion” of a face-to-face personal relationship between
individuals and media performers. Parasocial interaction theory was developed to describe
the interaction between media performers and media users, and parasocial interactions can
attract the audience to imagine themselves in the media program (Labrecque, 2014). A PSI
creates close familiarity with the performers as if they are intimate friends in the audience’s
daily lives (Berryman and Kavka, 2017; Xiang et al., 2016). Therefore, the media audience
“meets” Internet celebrities on a social media platform through observing their programs
(Kim et al., 2020; Utz, 2015). As these “meetings” become frequent and even daily, the
followers of the Internet celebrity imagine an illusory friendship formed from emotional Internet
bonds and affective attachment to the media performers (Qin, 2020; Xiang et al., 2016). celebrity
Furthermore, as the interpersonal interactions between the two parties move from a
superficial to intimate relationship, individuals share personal information voluntarily, such
characteristics
as feelings, opinions, comments and beliefs, with others based on the social presence
mechanism (Bao and Wang, 2021; Labrecque, 2014; Lai and Yang, 2015). Therefore, self-
disclosure produces a critical function in the parasocial interaction relationship process
beyond expert knowledge (Chung and Cho, 2017). 485
The purpose of this study is to explore how YouTube Internet celebrities demonstrate
their expertise, or self-exposed personality, to establish interactions with followers. We
investigate the understanding of parasocial interaction theory and attachment theory, and
examine how these Internet celebrities can increase followers’ impulse purchase behavior on
YouTube, in the production of content to generate monetary benefits. The emotional
connection includes parasocial interaction and attachment. Therefore, the causal relationship
between emotional connection and impulse purchase behavior is examined. This study
proposes an audience reaction model that includes a disclosure-type Internet celebrity driver
and a knowledge-type Internet celebrity driver and leads to impulse purchase behavior.
Impulse purchase behavior (IPB) is a non-planned buying action in which individuals buy
something without an intended shopping list and decide on the spot.
In interactive marketing, the effectiveness of social networks has been proved to be a
critical tactic (Akdevelioglu and Kara, 2020; Graham and Wilder, 2020; Wang, 2021). In
2020, YouTube was the world’s second most visited website (after Google), with an average
user visit time of 23 min (SimilarWeb, 2020). YouTube allows users to create and upload
videos that are seen and shared by hundreds of millions of viewers (Hwang and Zhang,
2018; Rosaen and Dibble, 2016), because YouTube provides a space to discuss products and
brands given the interaction of social media and social networks (Bi et al., 2019; Corr^ea et al.,
2020). YouTube Internet celebrities (YouTubers) are the site’s most active and well-known
users, who use their interactions to “influence words, goals, and activities on the platform”
(Bao and Wang, 2021; Corr^ea et al., 2020; Ferchaud et al., 2018). YouTube is a social platform
on which frequent brand interactions with consumers, including knowledge seeking and
self-identification, are one of the factors that build the affection of followers for an Internet
celebrity (Al-Emadi and Ben Yahia, 2020; Corr^ea et al., 2020; Qin, 2020). However, few
studies have empirically examined YouTube’s “subscribed” characteristics. Therefore, in
this study the YouTube platform is used to discuss the cause-and-effect relationship among
the characteristics of Internet celebrity, parasocial relationship and impulse purchase
behavior.

Literature review and development of hypotheses

Previous studies of YouTube Internet celebrities investigated followers’ reactions to these
celebrities to enhance the advertisement effect of intense parasocial interaction (Bi et al., 2019;
Corr^ea et al., 2020; Lee and Watkins, 2016). YouTube Internet celebrities have been deemed
content creators who encourage viewers to follow them as they perform in YouTuber-
generated content (Smith, 2016). Previously, celebrities became famous as singers or movie
stars, but today, YouTube Internet celebrities arises from the blurring of performers’ private
and public lives, and a sense of realness that comes from a perceived amateur performer
(Corr^ea et al., 2020; Jerslev, 2016). On YouTube platform, YouTubers are vloggers (Internet
celebrities) who share their video blogging (vlogs) in the blog format (Bi et al., 2019; Gong and
Li, 2017; Lai and Yang, 2015). Berryman and Kavka (2017) operated the practice process of
intimate access to a YouTube Internet celebrity’s life and the carefully planning of their role in
facilitating close relationships with followers.
JRIM Previous literature highlighted the specific way in which consumer engagement was
15,3 developed and increased through expertise-knowledge (EK) and/or self-disclosure (SD)
Internet celebrities to affect social media marketing activities and lead followers to subscribe
to YouTube channels and purchase products (Bao and Wang, 2021; Corr^ea et al., 2020;
Ferchaud et al., 2018; Gong and Li, 2017; Lee and Watkins, 2016). Many YouTubers who are
expert knowledge Internet celebrities post videos to teach their followers how to use certain
products, such as cosmetics, or to cook food, even combined toys are addressed. Expertise-
486 knowledge Internet celebrities start from a professional perspective and accumulate public
credibility, which develops cognitive reactions to produce good brand images and increases
consumers’ demand for products (Bao and Wang, 2021; Gong and Li, 2017). An expertise-
knowledge Internet celebrity increases the authenticity of the content, thus attracting
followers and generating more interest in the content (Lee and Watkins, 2016). Therefore,
expertise-knowledge is a critical factor in the formation of an interpersonal relationship. By
following an expert’s suggestion, followers tend to agree and are willing to change their
original attitudes (Ferchaud et al., 2018; Yuan et al., 2016). Therefore, YouTube viewers are
more likely to interact with Internet celebrities who are professionals with commodities,
companies, and environments in a certain field.
Alternatively, self-disclosure Internet celebrities share their daily lives, hobbies and pets,
personal experiences, private affairs, individual social lives, etc., on YouTube (Jerslev, 2016).
Self-disclosure means the Internet celebrity shares his or her feelings, opinions, comments,
daily life, private affairs and relationship status voluntarily with followers (Berryman and
Kavka, 2017; Utz, 2015). These celebrities focus on physical attractiveness, homophile
attitude and pleasing appeal (Smith, 2016; Xiang et al., 2016). Through the celebrity’s self-
disclosure in videos and posts on YouTube, followers feel similar feelings and intimacy with
the celebrities (Gong and Li, 2017; Welbourne and Grant, 2016). In this space, viewers can
interact with YouTubers and try to sort doubts about a brand advertised or simply propose
new experiments (Ferchaud et al., 2018).
This study provides the opportunity to compare this study’s findings to existing literature
and forms another contribution of this study. This study included self-disclosure Internet
celebrities who differ from traditional expertise-knowledge Internet celebrities who mainly
emphasize knowledge persuasion (Ferchaud et al., 2018; Qin, 2020; Utz, 2015). Self-disclosure
Internet celebrities who share personal experiences, practical daily living, and inner
emotional responses with subscribers create emotional bonds with them, which lead to
followers’ impulsive purchase behavior (Berryman and Kavka, 2017; Welbourne and Grant,
2016). This is attributed to the idea that mood-related features of Internet celebrities can
strengthen the cognitive and affective responses of image-sharing and affect impulsive
buying behavior (Styv
en et al., 2017; Utz, 2015). However, researchers have rarely empirically
examined the mediation of YouTube parasocial relationships. Therefore, this study searched
for evidence of mediating effects of parasocial relationships between Internet celebrity
characteristics and followers’ impulse purchase behavior on the YouTube platform. The
framework of the research is shown in Figure 1.

Relationships between self-disclosure and parasocial interaction and attachment

The self-disclosure process develops intimacy, and an individual expresses his or her
innermost senses and opinions to a partner deemed close, supportive, and worthy of
confidence (Ferchaud et al., 2018). Intimacy is considered a social closeness status produced
from private disclosures between two parties (Mal€ar et al., 2011). Furthermore, intimacy is
considered an individual and subjective feeling of links caused by an interpersonal and
reciprocal procedure of self-disclosure between close partners (Chen et al., 2019). Therefore,
audiences develop an attachment attitude toward an Internet celebrity, which indicates the
 Disclosure-typed
Internet Celebrity

Self- H1a Parasocial

Disclosure H4

H1b Impulse
H2a H3 Purchase
- Behavior
H5
Knowledge H2b

(H6: )
Knowledge-typed
Internet Celebrity
characteristics
celebrity
Internet

487

The framework of the

Figure 1.

research
JRIM likelihood of maintaining access to the celebrity (Gong and Li, 2017). The attachment sends a
15,3 signal on Facebook, Instagram, or email to transmit a message through a post or video to
confirm the emotional dependency on the Internet celebrity over time (Kim et al., 2020). Thus,
self-disclosure promises an imposing interpersonal action that affects others’ attachment
(Chen et al., 2019). Thus, the following hypothesis is proposed:
H1a. The self-disclosure of a YouTube Internet celebrity positively influences followers’
488 attachment to the Internet celebrity.
A celebrity’s self-disclosure enhances followers to feel his or her social presence (Kim and
Altmann, 2017). A close, usually highly involved, and deeply confessional readme are critical
characteristics of the self-disclosure of Internet celebrities (Chung and Cho, 2017). Intimate
self-disclosure allows viewers to penetrate the “public persona” of the individual and see the
Internet celebrity’s “innermost self” (Gong and Li, 2017). Followers want to exchange
information with the Internet celebrity on social media and make it easy to promote the
interaction process. Therefore, when a celebrity shares his or her daily life and actively
communicates about his or her experiences, followers perceive the social presence and then
an enhanced parasocial interaction relationship with him or her (Corr^ea et al., 2020; Kim et al.,
2020), because the nature of the self-disclosing interaction enhances perceived social
closeness and social bonds, therefore fostering a parasocial relationship (Ferchaud et al.,
2018). Thus, the following hypothesis is proposed:
H1b. The self-disclosure of a YouTube Internet celebrity positively influences the
parasocial interaction relationship between followers and the YouTube Internet
celebrity.

Relationships between expert knowledge and parasocial interaction and attachment

As Internet celebrities are recognized as having expertise-knowledge, audiences have
perceived and been persuaded that the endorsement makes effective claims (Ferchaud et al.,
2018). Followers then worship these Internet celebrities such that the followers want to
become the celebrity (Yue, 2010). Followers perceive the extent of the celebrity’s expertise-
knowledge and appreciate the Internet celebrity’s recommendations, which then creates idol
worship based on the mental health theory proposed by Yuan et al. (2016). This theory
normally refers to the image of idol worship in which followers develop an attachment and
then try to transmit the identical attributes from the idol to themselves (Cole and Leets, 1999).
Thus, an individual obtains a certain effective social skill by imitating other behavior which
enables him or her to reinforce the regulation of actions by observing others being
encouraged or punished for behaviors based on the social learning theory. Followers imitate
behaviors by considering their idols successful models; however, the worship and role
modeling are confounded (Yue, 2010). In short, an Internet celebrity has expert knowledge
and achievements in a specific field, which promotes followers to worship or to appreciate the
celebrities’ success and then imitate him or her and become attached emotionally (Lee and
Watkins, 2016). Thus, the following hypothesis is proposed:
H2a. The expertise-knowledge of a YouTube Internet celebrity positively influences
followers’ attachment to him or her.
Recent research in source credibility theory indicated that messages with credible resources
are considered more persuasive and more convincing than other messages, thus affecting
cognition, thoughts, opinions and attitudes (Jerslev, 2016; Ohanian, 1991; Styv
en et al., 2017).
As Internet celebrities have expertise-knowledge, their audiences perceive a positive
influence on product/service credibility from the celebrities’ persuasion effect (Yuan et al.,
2016). Media audiences of Internet celebrities perceive stronger source credibility from the
professional celebrity, and then, the audiences share messages and post them on social media Internet
(Graham and Wilder, 2020; Ohanian, 1991). For example, social media users can share celebrity
positive opinions, explain details, add comments and answer questions in messages
exchanged on Instagram, YouTube or Facebook because the relationship between the
characteristics
Internet celebrity and the followers is a close relationship, not an ephemeral or transient
relationship (Vander Schee et al., 2020). Therefore, frequent information exchange promotes a
parasocial interaction relationship between a YouTube Internet celebrity and his or her
audience, and the celebrity’s expertise-knowledge positively affects the formation of a 489
parasocial interaction relationship (Xiang et al., 2016). Thus, the following hypothesis is
proposed:
H2b. The expertise-knowledge of a YouTube Internet celebrity positively influences the
parasocial interaction relationship between followers and him or her.

Relationship between attachment and parasocial interaction

Attachment theory suggests that followers are more likely to form parasocial interaction
relationships with an Internet celebrity due to emotional bonds (Hinson et al., 2019; Rosaen
and Dibble, 2016; Stever, 2013). Every movement of an Internet celebrity will let followers
think about the bright light and attachment to each other. Followers want to imitate the
various behaviors of an Internet celebrity in their lives (Moulard et al., 2015). Attachment is
related to parasocial behavior because anxious-ambivalents, avoidants and secures from
attachment are most likely to form parasocial bonds which will engage individuals in
parasocial interaction through homophily (Cole and Letts, 1999; Kim and Altmann, 2017).
Followers imagine the Internet celebrity as a soulmate to worship and appreciate, and they
have an intense interest in the celebrity’s personal life, and thus, form specific parasocial
bonds (Jerslev, 2016; Moulard et al., 2015). To achieve the fulfillment that followers lack in
their daily lives and add a sense of purpose and excitement, followers form parasocial
interactions (Cole and Leets, 1999). Thus, the following hypothesis is proposed:
H3. Followers’ attachment to a YouTube Internet celebrity positively influences
parasocial interactions.

Relationship between attachment and impulse buying behavior

Attachment extends beyond people vs. people relations and has been described as emotion
overload with specific goals restricted between an individual and certain events, tasks or
objects (Corr^ea et al., 2020; Thomson, 2006). Followers develop attachment bonds to a
celebrity on behalf of favorite or special affairs based on attachment theory and self-
determination theory (Mal€ar et al., 2011; Thomson, 2006). The outcomes of the attachments
suggest that followers are convinced by a celebrity’s endorsement, and that produces inner
persuasion to accept products the celebrity recommends (Styv
en et al., 2017; Susarla et al.,
2016). Trust is a critical positive result of high celebrity attachment, and trust can produce
impulse buying behavior (Bao and Wang, 2021). Thus, based on impulse buying behavior
and the celebrity–follower relationship framework, the following hypothesis is proposed:
H4. Followers’ attachment to a YouTube Internet celebrity positively influences their
impulse buying behavior.

Relationship between parasocial interaction and impulse buying behavior

Parasocial interactions also affect followers’ impulse purchase behavior significantly because
an Internet celebrity with expert knowledge and/or self-disclosure, who has access to a
follower’s life, attracts followers who imitate the behavior of the Internet celebrity, similar to a
JRIM shopping program or celebrity (Corr^ea et al., 2020). Followers then imitate the style of the
15,3 Internet celebrity, follow recommendations in parasocial interactions, and finally, promote
impulse purchasing (Bao and Wang, 2021). On social media such as YouTube and Instagram,
users develop imagined intimacy and a parasocial interaction relationship with Internet
celebrities (Ferchaud et al., 2018). An Internet celebrity then develops an intimate relationship
with hosts who then persuade audiences, who make more unplanned purchases of
commodities (Hwang and Zhang, 2018; Munnukka et al., 2019). Since that a YouTuber’s
490 communication skills and personality attributes encourage audiences to develop and
preserve a parasocial interaction relationship with the provider, and this relationship
persuades followers to impulsively purchase products (Ferchaud et al., 2018; Lee and
Watkins, 2016). Therefore, the following hypothesis is proposed:
H5. Followers’ parasocial interaction with a YouTube Internet celebrity positively
influences their impulse buying behavior on social media.

Attachment and parasocial interaction mediate the link between Internet celebrity
characteristics and followers’ impulse purchase behavior
In this study, the causal framework is Internet celebrity characteristics (attachment and
parasocial interaction) and followers’ impulse purchase behavior, in which attachment and
parasocial interaction are deemed as mediation mechanisms. A self-disclosure Internet
celebrity develops intimacy through sharing personal experiences, practices of daily living,
and inner emotional responses; followers then develop an attachment attitude toward the
Internet celebrity (Berryman and Kavka, 2017). Differently, an expert knowledge Internet
celebrity has professional skills and achievements in a specific field, which promote followers
to worship or to appreciate the celebrity’s success, imitate him or her, and become attached
emotionally (Gong and Li, 2017). Therefore, the attachments suggest that followers are
convinced by a celebrity’s endorsement, and that produces followers’ impulse purchase
behavior (Labrecque, 2014; Susarla et al., 2016). Similarly, when a self-disclosure celebrity
shares individual daily life experiences and communicates vigorously with related
experiences, followers perceive the social presence and then an enhanced parasocial
interaction with the performer (Lee and Watkins, 2016). Followers of Internet celebrities with
expertise-knowledge perceive a positive source of credibility from the professional celebrity
and then share messages to influence the formation of the parasocial interaction (Xiang et al.,
2016). Therefore, the parasocial interaction with an Internet celebrity with expertise-
knowledge and/or self-disclosure significantly influences followers’ impulse purchase
behavior. Thus, the following hypothesis is proposed:
H6. Attachment and parasocial interaction mediate the link between YouTube Internet
celebrity characteristics (i.e. self-disclosure and expertise-knowledge Internet
celebrity) and followers’ impulse purchase behavior.

Research method
Measurement
Self-disclosure had nine measurable indicators within three latent constructs: honesty,
amount and depth (Gruzd and Hern
andez-Garc
ı, 2018). Expertise-knowledge had nine
measurable indicators within three latent constructs: expert, experience, and knowledge
(Yuan et al., 2016). Attachment had nine measurable indicators within three latent constructs:
relatedness, satisfaction and commitment (Oberecker and Diamantopoulos, 2011). Parasocial
interaction had nine measurable indicators within three latent constructs: cognition,
similarity and attraction (Yuan et al., 2016). Impulse purchase behavior had six measurable
indicators within two latent constructs: impulsive urge and lost self-regulation (Xiang et al., Internet
2016). This study included 500 respondents due to budget constraints, given that there were celebrity
39 items on the questionnaires, had ten times the number of measurement items (390), and
confirmed the survey was effective (Kline, 2011). We employed structural equation modelling
characteristics
(SEM) and used SPSS-AMOS 25.0 version software (J€oreskog and S€orbom, 1993).

Sampling procedure 491

This study collected data via an online survey from respondents amongst YouTube users
who are Internet celebrity followers in Taiwan. This study adopted quota sampling to
confirm the representativeness of the samples which the statistical features of the samples
meet the characteristic demographics of population. The population was separated into
different YouTube user sub-groups according to the 2018 population statistics for Taiwan.
The population data were classified by age and gender, the ordinary demographics that were
deemed as segmentation variables. This study determined what ratio of the population had
engaged in YouTube use according to gender and age from a survey of Taiwan’s online users
conducted by the Chuangshi Market Research Consulting Company Research Survey in 2018
(YouTube Pulse and Ipsos, 2018). As for the demographic characteristics of population, i.e.
YouTube total users in Taiwan, men accounted for 2.89 million persons (51%), and women
accounted for 2.78 million persons (49%). As for age, those 15–19 years totaled 1.19 million
persons (21%), 20–24 years totaled 1.42 million persons (25%), 25–29 years totaled 1.59
million persons (28%), and 30–34 years totaled 1.47 million persons (26%) (see Table 1). Note
that individuals younger than 15 year-olds were excluded because their income did not allow
them to purchase expensive products. Similarly, we follow the study of Xiang et al. (2016),
individuals elder than 35 years were excluded because the followers of Internet celebrities are
mainly young generations (under 35 year-olds).
This study designed a sample structure and decided how many samples were needed for
each cell. Therefore, we designed the dispatch of the 500 samples as follows. The planned
sample structure between men and women was 255 (51%) and 245 (49%), respectively. As for
age, the planned sample included 105 who were 15–19 years old (21%), 125 who were 20–24
years old (25%), 140 who were 25–29 years old (28%), and 130 who were 30–34 years
old (26%).
The study collected 500 effective samples, categorized by gender, age, and region in
Spring 2020. In Table 1, there were 253 male practical samples (50.6%) and 247 female
samples (49.4%). Among the actual participants, 111 were 15–19 years (22.2%), 128 were
20–24 years (25.6%), 135 were 25–29 years (27.0%) and 126 were 30–34 years (25.2%).
Empirical samples were also divided by region: northern (298; 59.6%), central (92; 18.4%),
southern (97; 19.4%) and eastern (13; 2.6%).
Additionally, we investigate if the random distribution of participants to different impulse
buying behavior was fine. We conduct one-way analysis of variance (ANOVA) analyses and
the result showed that impulse buying behavior did not differ significantly regarding
demographic features. The ANOVA is to examine whether the demographic characteristics
affected the outcome (impulse buying behavior) and to present the sample as representative
or not. This study chose gender, age, and region as the influencing variables. The p values of
0.700, 0.238 and 0.239, respectively (p > 0.05), for gender, area and region were not statistically
significant; impulse purchase behavior was not affected.

Common method variance issue

Regarding common method variance (CMV) issue, this study avoided the risk of spurious
relations amongst dependent and independent variables through introducing the cover
JRIM Demographics characteristics of population Items Popolation Percent
15,3
Gender Male 2,893,230 51
Female 2,779,770 49
Age Under 19 years old 1,194,440 21
20–24 years old 1,415,980 25
25–29 years old 1,589,250 28
492 30–34 years old 1,473,330 26
Total – 5,673,000 100

Planned sample structure Male Female Total

Under 19 years old 54 51 105 (21%)

20–24 years old 66 59 125 (25%)
25–29 years old 71 69 140 (28%)
30–34 years old 64 66 130 (26%)
Total 255 (51%) 245 (49%) 500 (100%)

500 empirical Percentage Cumulative

samples frequency Category Count (%) percentage (%)

Gender Male 253 50.6 50.6

Female 247 49.4 100
Age Under 19 years old 111 22.2 22.2
20–24 years old 128 25.6 47.8
25–29 years old 135 27.0 74.8
30–34 years old 126 25.2 100
Region Northern Taiwan 298 59.6 59.6
Central Taiwan 92 18.4 78.0
Southern Taiwan 97 19.4 97.4
Table 1. Eastern Taiwan and 13 2.6 100
Population, planned islands district
sample structure, and Total – 500 100 100
500 empirical samples Source(s): 2018 Taiwan’s “YouTube usage behavior survey” conducted by YouTube Pulse and Ipsos

picture at the interval between them. With Harman’s one-factor test of unrotated method, four
factors with eigenvalues were found to be larger than unity (rather than a single factor) within
42 items. These four factors explained 75.72% of the whole variance. None was obvious
because the percentage of explained variance regarding the first factor, 0.4785, did not exceed
0.5 (Podsakoff et al., 2003). Therefore, CMV was not a serious problem in this study.

Empirical result and testing of the hypotheses

Reliability and validity analysis
Cronbach’s alpha of 14 constructs of self-disclosure, expertise-knowledge, parasocial
interaction, attachment and impulse purchase behavior from 500 samples were 0.777,
0.853, 0.821, 0.876 and 0.865 that was larger than 0.7 and revealed acceptable internal
consistency (Bag€ozzi and Yi, 1988). We calculated values of composite reliability (CR) for self-
disclosure, expertise-knowledge, parasocial interaction, attachment and impulse purchase
behavior are 0.830, 0.889, 0.788, 0.887 and 0.845, respectively. All of the CR values were over
0.6, surpassing the criteria of reliability analysis, and it concluded that this test illustrates
internal consistency (Fornell and Larcker, 1981). This study computed the average variance
extracted (AVE) of self-disclosure, expertise-knowledge, parasocial interaction, attachment,
and impulse purchase behavior as 0.622, 0.728, 0.555, 0.725 and 0.732, respectively. This Internet
study found that AVE were greater than 0.5 and revealed that five variables properly celebrity
measured what they meant and stood convergent validity (Fornell and Larcker, 1981) (see
Table 2).
characteristics
Furthermore, the variance inflation factor (VIF) values ranged from 1.296 to 2.575 that
were lower than three, represented that multicollinearity problem were not exists (Hair et al.,
2019). The coefficient of determination (R2) ranged from 0.205 to 0.364 that expressed a
stronger explanatory power in this study given that R2 was considered little, medium, or huge 493
values of 0.02, 0.13, and 0.26, respectively (Hair et al., 2019). We calculated that effect sizes (f2),
f2 5 [R2/(1 R2)], ranged from 0.258 to 0.572, which belonged to medium and large effects.
Since that 0.02, 0.15 and 0.35 was deemed as little, medium and huge effects in evaluating the
intensity of the relationship among related variables (see Table 3).

Results of structural equations modeling

This study obtained χ 2/df was 4.479, that fit a criteria 2 < X < 5 (Anderson and Weitz, 1992).
This study calculated GFI (0.928), AGFI (0.882), CFI (0.948), IFI (0.949), NFI (0.935) that was
larger than 0.9 to show the validity was hold (Bag€ozzi and Yi, 1988). Again, RMSR was 0.044
(less than 0.050) and RMESA was 0.084 (smaller than 0.100) that indicated a proper fitness.
The self-disclosure Internet celebrity had positive significant effect on attachment (H1a:
β1 5 0.525, t-value 5 7.148); however, self-disclosure had not positive significant effect on
parasocial interaction (H1b: β2 5 0.037, t-value 5 0.975). The expertise-knowledge Internet
celebrity had positive significant effect on attachment (H2a: β3 5 0.404, t-value 5 6.042); and
expertise-knowledge had positive significant effect on parasocial interaction (H2b: β4 5 0.231,
t-value 5 5.710). Followers’ attachment to an Internet celebrity had positive significant effect
on parasocial interaction (H3: β3 5 0.231, t-value 5 5.710); and attachment had positive
significant effect on their impulse purchase behavior (H4: β4 5 0.666, t-value 5 2.111).
Followers’ parasocial interaction with Internet celebrity had positive significant effect on
their impulse purchase behavior. (H5: β5 5 0.624, t-value 5 4.235) (see Table 3).

Cronbach’s Average
α if Error- Composite variance
Cronbach’s construct term reliability extracted
Variables Constructs α deleted Loading estimated (CR) (AVE)

Self- Amount 0.777 0.762 0.643 0.371 0.830 0.622

disclosure Depth 0.581 0.852 0.212
(SD) Honesty 0.736 0.718 0.422
Expertise- Expert 0.853 0.836 0.755 0.309 0.889 0.728
knowledge Experience 0.758 0.845 0.218
(EK) knowledge 0.786 0.819 0.203
Parasocial Cognitive 0.821 0.681 0.675 0.384 0.788 0.555
interaction Similarity 0.816 0.764 0.306
(PSI) Attraction 0.759 0.631 0.464
Attachment Relatedness 0.876 0.817 0.866 0.210 0.887 0.725
(AT) Satisfaction 0.855 0.722 0.329
Commitment 0.800 0.861 0.226
Impulse Impulse 0.865 – 0.900 0.230 0.845 0.732
purchase urge Table 2.
behavior Lost self- – 0.834 0.322 Reliability analysis of
(IPB) regulation each construct
JRIM Hypothesized path Coefficient T-value VIF R2 f2
15,3
H1a: Self-disclosure →Attachment β1a 5 0.525 7.148*** 1.296 0.250 0.333
H1b: Self-disclosure →Parasocial interaction β1b 5 0.037 0.975 1.493 0.290 0.408
H2a: Expertise-knowledge →Attachment β2a 5 0.404 6.042*** 1.296 0.205 0.258
H2b: Expertise-knowledge →Parasocial interaction β2b 5 0.231 5.710*** 1.408 0.364 0.572
H3: Attachment →Parasocial interaction β3 5 0.710 17.773*** 1.451 0.611 1.571
494 H4: Attachment → Impulse purchase behavior β4 5 0.666 2.111* 2.575 0.219 0.280
H5: Parasocial interaction → Impulse purchase behavior β5 5 0.624 4.235*** 2.575 0.231 0.300

Direct Total
Path effect Indirect effect effect Ranking Aggregation

SD→IPB SD → AT → IPB 0 0.525*0.666 0.3497 1 0.6054

SD → AT → PSI → IPB 0 0.525*0.710*0.624 0.2326 3
SD → PSI → IPB 0 0.037*0.624 0.0231 6
EK→IPB EK → AT → IPB 0 0.404*0.666 0.2691 2 0.5902
EK → AT → PSI → IPB 0 0.404*0.710*0.624 0.1790 4
EK → PSI → IPB 0 0.231*0.624 0.1441 5
Table 3.
Results of research Note(s): Based on one-tailed test: for t-value greater than 1.96 of 0.05 significant level (*); for t-value greater
hypothesis (SEM) and than 2.33 of 0.00 significant level (**); for t-value greater than 2.58 of 0.001 significant level (***). χ 2/df 5 4.479,
direct and indirect GFI 5 0.928, AGFI 5 0.882, CFI 5 0.948, IFI 5 0.949, NFI 5 0.935, RMSR 5 0.044, and RMSEA 5 0.084. SD is
effects of self-disclosure, EK is Expertise-knowledge, PSI is parasocial interaction, AT is attachment, IPB is impulse
different paths purchase behavior

Mediating effect analysis

This study employed the Bootstrap method suggested by Lau and Cheung (2012) to estimate
the mediating effect. In case 1 (SD-IP), the confidence interval (0.091–0.309) for indirect effects
did not contain zero, and had a significant effect (p < 0.05). The confidence interval ( 0.327–
0.057) of the direct effect contains zero, indicating that the direct effect was not significant.
The result was the same as case 1, which showed that PSI and AT had a full mediating effect
between SD-IP. In case 2 (EK-IP), the confidence interval (0.065–0.559) for indirect effects did
not contain zero, and had a significant effect (p < 0.05). The confidence interval ( 0.245–
0.229) of the direct effect contained zero, indicating that the direct effect was not
significant. It showed that PSI and AT had a full mediating effect between EK-IP (see
Table 4).

Empirical analysis of separate the Internet celebrities to different types

This study separated Internet celebrities into two types: “self-disclosure” and “expertise-
knowledge.” The disclosure-typed Internet celebrity type had 244 samples and the
knowledge-typed Internet celebrity had 256 samples. One-way ANOVA indicated that
there was a significant difference between two Internet celebrity types and impulse purchase
behavior (F 5 9.052; p 5 0.020); thus, we used different types of Internet celebrities in SEM
model and found different and opposite results in Table 5. The main route in disclosure-typed
Internet celebrity was that self-disclosure affect impulsive purchase behavior through
parasocial interaction and attachment; oppositely, the main route in knowledge-typed
Internet celebrity was that expertise-knowledge affected impulsive purchase behavior
through parasocial interaction and attachment.
Additionally, we conduct one-way analysis of variance (ANOVA) analyses again as post
hoc tests for two Internet celebrity types and impulse buying behavior. As to self-disclosure
Effects Contents Estimates p-value Confidence interval Mediation effects
Internet
celebrity
Case1: SD → PSI → IPB characteristics
SD → AT → IPB
Indirect effect SD → PSI → IPB 0.190 0.000 0.104–0.309 Full mediating
SD → AT → IPB 0.148 0.001 0.091–0.232
Direct effect SD → IPB 0.113 0.195 0.327–0.057
SD → PSI 0.462 0.001 0.330–0.586 495
PSI → IPB 0.411 0.001 0.203–0.612
SD → AT 0.444 0.001 0.308–0.578
AT → IPB 0.332 0.001 0.221–0.452
Total effect SD → IPB 0.224 0.003 0.077–0.354
Case2: EK→ PSI → IPB
EK → AT → IPB
Indirect effect EK → PSI → IPB 0.195 0.000 0.144–0.559 Full mediating
EK → AT → IPB 0.116 0.000 0.065–0.202
Direct effect EK → IPB 0.007 0.993 0.245–0.229
EK → PSI 0.474 0.001 0.326–0.615
PSI → IPB 0.411 0.001 0.203–0.612
EK → AT 0.349 0.001 0.199–0.492 Table 4.
AT → IPB 0.332 0.001 0.221–0.45 Empirical result of
Total effect EK → IPB 0.318 0.002 0.162–0.452 serial mediation effects

Items The self-closure group The expertise-knowledge group

Estimate t-value p-value Estimate t-value p-value

SD → PSI β1 5 5.529 3.909 *** β1 5 0.208 1.385 –

SD → AT β2 5 10.793 3.266 *** β2 5 0.263 1.451 –
EK → PSI β3 5 3.256 3.069 – β3 5 0.792 8.643 ***
EK → AT β4 5 7.239 2.905 – β4 5 0.677 7.249 ***
PSI → IPB β5 5 2.850 4.642 *** β5 5 0.105 0.341 –
AT → IPB β6 5 1.926 3.460 *** β6 5 0.873 3.047 ***

χ2 df χ 2 =df
The social-disclosure group 178.16 67 2.66
The expertise-knowledge group 194.06 63 3.08
Difference 15.9 4 (>χ 20:05;1 5 3.84)
Note(s): SD is social-disclosure; EK is expertise-knowledge; PSI is parasocial interaction; AT is attachment;
IPB is impulse purchase behavior. Based on one-tailed test: for t-value greater than 1.96 of 0.05 significant level
(*); for t-value greater than 2.33 of 0.00 significant level (**); for t-value greater than 2.58 of 0.001 significant
level (***)
As for the social-disclosure group, the significant ratio of the study model is 50%; χ 2 =df 5 2.659; GFI 5 0.907; Table 5.
AGFI 5 0.854; CFI 5 0.955; NFI 5 0.930; IFI 5 0.955; RMSR 5 0.041; and RMSEA 5 0.083 Model comparison
As for the expertise-knowledge group, the significant ratio of the study model is 50%; χ 2 =df 5 3.080; between SD and EK
GFI 5 0.907; AGFI 5 0.845; CFI 5 0.933; NFI 5 0.906; IFI 5 0.934; RMSR 5 0.039; and RMSEA 5 0.090 groups

Internet celebrity, we found that the p-values of gender (p 5 0.784), age (p 5 0.304), and region
(p 5 0.116) were not significant in affecting impulse buying behavior (p > 0.05). The result
showed that impulse buying behavior did not differ significantly regarding demographic
features for self-disclosure Internet celebrity. As to expertise-knowledge Internet celebrity,
this study found that the p-values of gender (p 5 0.758), age (p 5 0.228) and region (p 5 0.077)
JRIM were not significant in affecting impulse buying behavior (p > 0.05). Similarly result was
15,3 found for expertise-knowledge Internet celebrity (see Table 6).

Discussion and conclusions

Based on the empirical results, the self-disclosure and expertise-knowledge of YouTube
Internet celebrities affected followers’ impulse purchase behavior through attachment and
496 parasocial interaction. The first path indicated that self-disclosure YouTube Internet
celebrities positively affected attachment, which mainly positively affected followers’
impulse purchase behavior. The secondary path showed that YouTube Internet celebrities
with expertise-knowledge positively affected attachment, which positively affected followers’
impulse purchase behavior. Furthermore, self-disclosure and expertise-knowledge YouTube
Internet celebrities influence followers’ impulse purchase behavior through the mediation of
attachment and parasocial interaction; the aggregation equaled 0.6054 and 0.5902,
respectively. This result showed that the mediation effect definitively played an important
role in parasocial relationships in interactive marketing.

Theoretical contribution
Regarding academic implications, first, this study separated two types of influence processes
and discriminated attachment and parasocial interaction as two kinds of intense emotions to
form impulse purchase behavior based on attachment theory and parasocial interaction
theory (Cole and Leets, 1999; Hinson et al., 2019; Labrecque, 2014; Lee and Watkins, 2016).
This study is different from those conducted by Hinson et al. (2019) and Rosaen and Dibble
(2016) that investigated a single variable, attachment. This study is also different from three
recent studies conducted by Chung and Cho (2017), Gong and Li (2017), and Xiang et al. (2016)
that investigated a single variable, parasocial relationship. Empirical results of this study
have a more comprehensive explanation mechanism for the parasocial relationship factor.
Thus, this study integrated attachment and parasocial interaction to verify a novel parasocial
relationship to mediate YouTube Internet celebrity characteristics and followers’ impulsive
purchasing. Empirical results showed that the psychological state of attachment with a
YouTube Internet celebrity was obvious: It was the critical influencing factor in parasocial
interactions with self-disclosure and expertise-knowledge YouTube Internet celebrities.
Therefore, the followers’ level of impulsive purchase behavior was obvious.

Impulse purchase behavior Gender Age Region

Total Internet celebrity

F-value 1.835 1.398 1.411
p-value 0.700 0.238 0.239
Testing Non-reject Non-reject Non-reject
Self-disclosure type
F-value 0.075 1.062 1.987
p-value 0.784 0.304 0.116
Testing Non-reject Non-reject Non-reject
Table 6.
ANOVA analysis of Expertise-knowledge type
impulse purchase F-value 0.095 1.457 2.312
behavior for gender, p-value 0.758 0.228 0.077
age, and region Testing Non-reject Non-reject Non-reject
 Second, this study integrated self-disclosure YouTube Internet celebrities and YouTube Internet
Internet celebrities with expert knowledge to explore the role of their parasocial relationships celebrity
in impulse purchase behavior. This study included the two types of Internet celebrities, which
is different from studies conducted by Lai and Yang (2015), Utz (2015) and Welbourne and
characteristics
Grant (2016) that explored a single type, the self-disclosure Internet celebrity, and studies
conducted by Hwang and Zhang (2018), Munnukka et al. (2019) and Yuan et al. (2016) that
investigated a single type, the expertise-knowledge Internet celebrity. This study then
provided a comparative discussion of the self-disclosure type and expertise-knowledge type 497
of YouTube Internet celebrities that affected followers’ impulse purchase behavior. Empirical
results showed that self-disclosure Internet celebrities who created an attachment with
emotional bonds between themselves and their followers then successfully influenced
impulsive purchases. This study filled the research gap.

Managerial implications
Regarding practical implications, social media marketers should proceed prudently with
market segmentation and choose the appropriate type of Internet celebrities who are suitable
for a product’s image to differentiate marketing (Al-Emadi and Ben Yahia, 2020; Qin, 2020).
The present study results provided recommendations for message providers for Internet
celebrities to benefit their interactive marketing strategies to effectively disseminate official
messages through different Internet celebrities.
For self-disclosure Internet celebrities, marketers can emphasize emotional consumption
and remarked as experimental marketing in personal cosmetics, private accessories and
clothing accessories. Marketers need to guide Internet celebrities to highlight their personal
characteristics; then followers will empathize in interactive marketing. Marketers must assist
Internet celebrities in a planned way and regularly provide various events or hold activities.
This strategy enables Internet celebrities to enrich their self-disclosure materials to maintain
their freshness and popularity on YouTube (Gong and Li, 2017). Marketers also need to
analyze the frequency and response of interactions between Internet celebrities and followers
to understand the true demands of targeted market.
For expertise-knowledge Internet celebrities, marketers need to guide and expand the
relevant learning experience of followers on social media to emphasize the celebrities’
credibility and online popularity (Kim et al., 2020). This strategy produces an obvious
attachment which enhances the purchase persuasion effect through emotional bonds and
promises an important interactive marketing strategy, which cannot be ignored, such as
educational commodities, health care supplies, and financial and insurance commodities.
Marketers must provide Internet celebrities with sufficient expertise-knowledge about the
product, i.e. product functions, ingredient attributes, brand value, etc.

Limitations and future research directions

This study had three limitations. First, YouTube social media platform was examined. If
other well-known social media platforms (e.g. Instagram and Facebook) are compared, then
the difference effect on social media can be seen. Conducting cross-country or cross-media
comparisons would be better. Second, the sample volume and the scope of the data collection
could be expanded. The statistics had recommended collecting more than 1,000 samples to
reduce the opportunity for statistical biases. Budget constraints limited the number to 500
samples, and a non-Taiwanese sample was not incorporated. Third, the population of this
study might be more accurate. The population to whom the questionnaire was administered
was the current Internet celebrity economy in Taiwan (Taiwan’s top 10 YouTubers). It is
more effective to provide practical marketing implications through actual market
segmentation and target products for a real target market.
JRIM References
15,3 Akdevelioglu, D. and Kara, S. (2020), “An international investigation of opinion leadership and social
media”, Journal of Research in Interactive Marketing, Vol. 14 No. 1, pp. 71-88, doi: 10.1108/JRIM-
11-2018-0155.
Al-Emadi, F.A. and Ben Yahia, I. (2020), “Ordinary celebrities related criteria to harvest fame and
influence on social media”, Journal of Research in Interactive Marketing, Vol. 14 No. 2,
pp. 195-213, doi: 10.1108/JRIM-02-2018-0031.
498
Anderson, E. and Weitz, B. (1992), “The use of pledges to build and sustain commitment in
distribution channels”, Journal of Marketing Research, Vol. 29 No. 1, pp. 18-34.
Bag€ozzi, R.P. and Yi, Y. (1988), “On the evaluation of structural equation models”, Journal of the
Academy of Marketing Science, Vol. 16 No. 1, pp. 74-94.
Bao, Z. and Wang, D. (2021), “Examining consumer participation on brand microblogs in China:
perspectives from elaboration likelihood model, commitment–trust theory and social presence”,
Journal of Research in Interactive Marketing, Vol. 15 No. 1, pp. 10-19, doi: 10.1108/JRIM-02-
2019-0027.
Berryman, R. and Kavka, M. (2017), “I guess a lot of people see me as a big sister or a friend: the role of
intimacy in the celebrification of beauty vloggers”, Journal of Gender Studies, Vol. 26 No. 3,
pp. 307-320, doi: 10.1080/09589236.2017.1288611.
Bi, N.C., Zhang, R. and Ha, L. (2019), “Does valence of product review matter?”, Journal of Research in
Interactive Marketing, Vol. 13 No. 1, pp. 79-95, doi: 10.1108/JRIM-04-2018-0049.
Chen, L., Hu, N., Shu, C. and Chen, X. (2019), “Adult attachment and self-disclosure on social
networking site: a content analysis of Sina Weibo”, Personality and Individual Differences,
Vol. 138 No. 1, pp. 96-105.
Chung, S. and Cho, H. (2017), “Fostering parasocial relationships with celebrities on social media:
implications for celebrity endorsement”, Psychology and Marketing, Vol. 34 No. 4, pp. 481-495.
Cole, T. and Leets, L. (1999), “Attachment styles and intimate television viewing: insecurely forming
relationships in a parasocial way”, Journal of Social and Personal Relations, Vol. 16 No. 4,
pp. 495-511, doi: 10.1177/0265407599164005.
Corr^ea, S.C.H., Soares, J.L., Christino, J.M.M., Gosling, M.d.S. and Gonçalves, C.A. (2020), “The influence
of YouTubers on followers’ use intention”, Journal of Research in Interactive Marketing, Vol. 14
No. 2, pp. 173-194, doi: 10.1108/JRIM-09-2019-0154.
Ferchaud, A., Grzeslo, J., Orme, S. and LaGroue, J. (2018), “Parasocial attributes and YouTube
personalities: exploring content trends across the most subscribed YouTube channels”,
Computers in Human Behavior, Vol. 80 No. 1, pp. 88-96.
Fornell, C. and Larcker, D.F. (1981), “Evaluating structural equation models with unobservable
variables and measurement error”, Journal of Marketing Research, Vol. 18 No. 1, pp. 39-50, doi:
10.2307/3151312.
Gong, W. and Li, X. (2017), “Engaging fans on microblog: the synthetic influence of parasocial
interaction and source characteristics on celebrity endorsement”, Psychology and Marketing,
Vol. 34 No. 2, pp. 720-732.
Graham, K.W. and Wilder, K.M. (2020), “Consumer-brand identity and online advertising message
elaboration: effect on attitudes, purchase intent and willingness to share”, Journal of Research in
Interactive Marketing, Vol. 14 No. 1, pp. 111-132, doi: 10.1108/JRIM-01-2019-0011.
Gruzd, A. and Hern
andez-Garc
ıa, A.
(2018), “Privacy concerns and self-disclosure in private and
public uses of social media”, Cyberpsychology, Behaviour, and Social Networking, Vol. 21 No. 7,
pp. 418-428.
Hair, J.F., Risher, J.J., Sarstedt, M. and Ringle, C.M. (2019), “When to use and how to report the results
of PLS-SEM”, European Business Review, Vol. 31 No. 1, pp. 2-24.
Hinson, R., Boateng, H., Renner, A. and Kosiba, J.P.B. (2019), “Antecedents and consequences of Internet
customer engagement on Facebook: an attachment theory perspective”, Journal of Research in
Interactive Marketing, Vol. 13 No. 2, pp. 204-226, doi: 10.1108/JRIM-04-2018-0059. celebrity
Hwang, K. and Zhang, Q. (2018), “Influence of parasocial relationship between digital celebrities and
characteristics
their followers on followers’ purchase and electronic word-of-mouth intentions, and persuasion
knowledge”, Computers in Human Behavior, Vol. 87 No. 1, pp. 155-173.
Jerslev, A. (2016), “Media times in the time of the microcelebrity: celebrification and the YouTuber
zoella”, International Journal of Communication, Vol. 10 No. 4, pp. 5233-5251, available at: http:// 499
refhub.elsevier.com/S0747-5632(17)30615-5/sref32.
J€oreskog, K.G. and S€orbom, D. (1993), LISREL 8: Structural Equation Modeling with the SIMPLIS
Command Language, Scientific Software International, New York, NY.
Kim, K. and Altmann, J. (2017), “Effect of homophily on network formation”, Communications in
Nonlinear Science and Numerical Simulation, Vol. 44 No. 3, pp. 482-494.
Kim, J.M., Kim, M. and Key, S. (2020), “When profile photos matter: the roles of reviewer profile photos
in the online review generation and consumption processes”, Journal of Research in Interactive
Marketing, Vol. 14 No. 4, pp. 391-412, doi: 10.1108/JRIM-10-2019-0163.
Kline, R. (2011), Principles and Practice of Structural Equation Modeling, 3rd ed., Guilford, New York,
NY, ISBN: 978-1-60623-876-9.
Labrecque, L.I. (2014), “Fostering consumer–brand relationships in social media environments: the
role of parasocial interaction”, Journal of Interactive Marketing, Vol. 28 No. 2, pp. 134-148, doi:
10.1016/j.intmar.2013.12.003.
Lai, C.Y. and Yang, H.L. (2015), “Determinants of individuals’ self-disclosure and instant
information sharing behavior in micro-blogging”, New Media and Society, Vol. 17 No. 9,
pp. 1454-1472.
Lau, R.S. and Cheung, G.W. (2012), “Estimating and comparing specific mediation effects in complex
latent variable models”, Organizational Research Methods, Vol. 15 No. 1, pp. 3-16.
Lee, J.E. and Watkins, B. (2016), “YouTube vloggers’ influence on consumer luxury brand perceptions
and intentions”, Journal of Business Research, Vol. 69 No. 12, pp. 5753-5760, doi: 10.1016/j.
jbusres.2016.04.171.
Mal€ar, L., Krohmer, H., Hoyer, W.D. and Nyffenegger, B. (2011), “Emotional brand attachment and
brand personality: the relative importance of the actual and the ideal self”, Journal of Marketing,
Vol. 75 No. 4, pp. 35-52.
Moulard, J.G., Garrity, C.P. and Rice, D.H. (2015), “What makes a human brand authentic? Identifying
the antecedents of celebrity authenticity”, Psychology and Marketing, Vol. 32 No. 2, pp. 173-186,
doi: 10.1002/mar.20771.
Munnukka, J., Maity, D., Reinikainen, H. and Luoma-aho, V. (2019), “‘Thanks for watching’. The
effectiveness of YouTube vlog endorsements”, Computers in Human Behaviour, Vol. 93 No. 2,
pp. 226-234.
Oberecker, E.M. and Diamantopoulos, A. (2011), “Consumers’ emotional bonds with foreign countries:
does consumer affinity affect behavioral intentions?” , Journal of International Marketing,
Vol. 19 No. 2, pp. 45-72.
Ohanian, R. (1991), “The impact of celebrity spokespersons’ perceived image on consumers’ intention
to purchase”, Journal of Advertising Research, Vol. 31 No. 1, pp. 46-54.
Ong, K., J€arvelin, K., Sanderson, M. and Scholer, F. (2017), “Using information scent to understand
mobile and desktop web search behavior”, 2017 ACM SIGIR ’17, Session 3A, August 07-11,
Shinjuku, Tokyo, Japan.
Podsakoff, P.M., MacKenzie, S.B., Lee, J.Y. and Podsakoff, N.P. (2003), “Common method biases in
Behavioral research: a critical review of the literature and recommended remedies”, Journal of
Applied Psychology, Vol. 88 No. 5, pp. 879-903.
JRIM Qin, Y.S. (2020), “Fostering brand–consumer interactions in social media: the role of social media uses
and gratifications”, Journal of Research in Interactive Marketing, Vol. 14 No. 3, pp. 337-354, doi:
15,3 10.1108/JRIM-08-2019-0138.
Rosaen, S.F. and Dibble, J.L. (2016), “Clarifying the role of attachment and social compensation on
parasocial relationships with television characters”, Communication Studies, Vol. 67 No. 2,
pp. 147-162, doi: 10.1080/10510974.2015.1121898.
SimilarWeb (2020), “2020 Digital trends”, available at: www.similarweb.com/corp/reports/2020-
500 digitaltrends-report/ (accessed 21 February 2020).
Smith, D.R. (2016), “Imagining others more complexly: celebrity and the ideology of fame among
YouTube’s ‘Nerdfighteria’”, Celebrity Studies, Vol. 7 No. 3, pp. 339-353, doi: 10.1080/19392397.
2015.1132174.
Stever, G.S. (2013), “Mediated vs parasocial relationships: an attachment perspective”, Journal of
Media Psychology, Vol. 17 No. 3, pp. 1-31.
Styv
en, M.E., Foster, T. and Wallstr€om, 
A. (2017), “Impulse buying tendencies among online shoppers
in Sweden”, Journal of Research in Interactive Marketing, Vol. 11 No. 4, pp. 416-431, doi: 10.1108/
JRIM-05-2016-0054.
Susarla, A., Oh, J.H. and Tan, Y. (2016), “Influentials, imitables, or susceptibles? Virality and word-of-
mouth conversations in online social networks”, Journal of Management Information Systems,
Vol. 33 No. 1, pp. 139-170.
Thomson, M. (2006), “Human brands: investigating antecedents to consumers’ strong attachments to
celebrities”, Journal of Marketing, Vol. 70 No. 3, pp. 104-119.
Utz, S. (2015), “The function of self-disclosure on social network sites: not only intimate, but also
positive and entertaining self-disclosures increase the feeling of connection”, Computers in
Human Behavior, Vol. 45 No. 1, pp. 1-10.
Vander Schee, B.A., Peltier, J. and Dahl, A.J. (2020), “Antecedent consumer factors, consequential
branding outcomes and measures of online consumer engagement: current research and future
directions”, Journal of Research in Interactive Marketing, Vol. 14 No. 2, pp. 239-268, doi: 10.1108/
JRIM-01-2020-0010.
Wang, C.L. (2021), “New frontiers and future directions in interactive marketing: inaugural Editorial”,
Journal of Research in Interactive Marketing, Vol. 15 No. 1, pp. 1-9, doi: 10.1108/JRIM-03-
2021-270.
Welbourne, D.J. and Grant, W.J. (2016), “Science communication on YouTube: factors that affect
channel and video popularity”, Public Understanding of Science, Vol. 25 No. 6, pp. 706-718.
Xiang, L., Zheng, X., Lee, M.K. and Zhao, D. (2016), “Exploring consumers’ impulse buying Behaviour
on social commerce platform: the role of parasocial interaction”, International Journal of
Information Management, Vol. 36 No. 3, pp. 333-347.
YouTube Pulse and Ipsos (2018), “2018 Taiwan’s YouTube usage behavior survey”, available at:
https://www.thinkwithgoogle.com/intl/zh-tw/consumer-insights/consumer/power-shoppers/.
Yuan, C.L., Kim, J. and Kim, S.J. (2016), “Parasocial relationship effects on customer equity in the social
media context”, Journal of Business Research, Vol. 69 No. 9, pp. 3795-3803.
Yue, X. (2010), “On idol transformation education: thoughts for helping young people to benefit from
idol worship”, Asian Journal of Counselling, Vol. 17 No. 1, pp. 23-40.

About the authors

Tser Yieth Chen is a Professor of Graduate Institute of International Business at the National Taipei
University in Taiwan. He completed his PhD degree in Management Science at National Chiao-Tung
University in Taiwan. His main research interests are customer behavior, marketing management,
service management, and performance management. His more than 30 articles have been published in
SSCI or SCI Journal such as: Asia Pacific Journal of Marketing and Logistics, Journal of Business and
Psychology, International Journal of Service Industry Management, Innovation: Organization and
Management, Service Industries Journal, Growth and Change, International Journal of Human Resource Internet
Management, Programs: Electronic Library and Information Systems, Energy Economics, Energy
Policy, European Journal of Operational Research, Journal of Operational Research Society, and celebrity
Applied Economics Letters, etc. Tser Yieth Chen is the corresponding author and can be contacted at: characteristics
[email protected]
Tsai Lien Yeh is a Professor of Department of International Business at the Ming-Chuan University
in Taiwan. He completed his PhD degree in Management at National Taiwan University of Science and
Technology in Taiwan. Her main research interests are marketing management, performance 501
management, and technology and innovation management. His more than 10 articles have been
published in in SSCI or SCI Journal such as: Innovation: Organization and Management, Asia Pacific
Journal of Marketing and Logistics, Service Industries Journal, International Journal of Service Industry
Management, Energy Policy, and International Journal of Sustainable Development and World
Ecology, etc.
Fang Yu Lee is a Master of Graduate Institute of International Business at the National Taipei
University in Taiwan.

For instructions on how to order reprints of this article, please visit our website:
www.emeraldgrouppublishing.com/licensing/reprints.htm
Or contact us for further details: [email protected]
Reproduced with permission of copyright owner. Further
reproduction prohibited without permission.