Journal of Retailing and Consumer Services xxx (xxxx) xxxx

Contents lists available at ScienceDirect

Journal of Retailing and Consumer Services

journal homepage: www.elsevier.com/locate/jretconser

Instagram and YouTube bloggers promote it, why should I buy? How
credibility and parasocial interaction influence purchase intentions
Karina Sokolova , Hajer Kefi
⁎

PSB: Paris School of Business, Digital Data Design, Chair D3, 59 rue Nationale, Paris, France

ARTICLE INFO ABSTRACT

Keywords: Nowadays social network influencers play an important role in marketing by introducing products to their
Parasocial interaction audience. In this article, we investigate the persuasion cues related to beauty and fashion influencers present on
Credibility YouTube and Instagram. More precisely, we investigate how the para-social interaction (PSI) the audience
Purchase Intention creates with the online influencer, along with their perceived credibility, are related to the purchase intention
Social networks
and how they are, in turn, related to the social and physical attractiveness and attitude homophily. We base our
Influencer
Beauty Brands
research on four beauty influencers popular in France and control our results by the age of the participants and
Persuasion by the influencer. We find that attitude homophily is positively related to PSI but, surprisingly, the physical
Communicator attractiveness shows negative relationship or no evidence of relationship. Both credibility of the influencers and
PSI exhibit significant and positive relationships to purchase intention.

1. Introduction refer to them as influencers (Susarla et al., 2016; Zhu et al., 2015;
Uzunoğlu and Kip, 2014).
Recently, there has been a spate of interest in online social media, Influencers are present in all sectors: health and fitness, fashion and
such as Facebook, Twitter, YouTube, Instagram and others (Knoll, beauty, food, high-tech and other (Raggatt et al., 2018; Klassen et al.,
2016; Dwivedi et al., 2018; Shiau et al., 2017). Presence on social 2018). YouTube, Instagram and Facebook bloggers often present the
networks has become a crucial part of marketing strategies: the pro- products they have tested providing their opinion or promoting them
motion of products online via business accounts and advertising on online to other users. The posts of influencers on Instagram usually take
social networking platforms has proven to be efficient (De Vries and the form of an image or video enriched with embedded content and a
Carlson, 2014; Gensler et al., 2013). textual description. The popularity of bloggers and their UGC has
In addition to the branded posts and fan pages found on social brought a new form of marketing called ‘influencer marketing’ where
media, users actively generate and publish multimedia content in- brands collaborate with influencers to promote their products (De
cluding their opinions on brands and products. Such content, also Veirman et al., 2017).
known as User Generated Content, has proven to be more popular and Brands and businesses are increasingly interested in finding ways to
effective than professional advertisements (Welbourne and Grant, leverage ‘popular figures’ as brand ambassadors on social media, and
2016; Aral et al., 2013; Lipizzi et al., 2015). The traditional word-of- those are not necessarily famous singers, actors or athletes. As recently
mouth (WOM) has moved online becoming an electronic word-of- reported by an industrial study conducted by Business Insider
mouth (eWOM) (Chang et al., 2015; Hennig-Thurau et al., 2004; (Gallagher, 2018), it is critical for many brands to find and manage
Alalwan et al., 2017; Kapoor et al., 2018). Thus, brands and customers relationships with social media influencers. According to the same
engage in collaborative processes of knowledge exchange and product- study, investments in this area are poised to reach between $5 billion
related information sharing (Merz et al., 2018; Shi et al., 2016). and $10 billion in 2022 (that is a five-year compound annual growth
Certain users can be active creators of online content: personal rate of 38%), especially in the sectors of fashion, beauty, parenting and
stories, ideas, reviews, opinions, feelings, emotions, etc., intended for a tourism. Whereas, many other industrial studies are reporting similar
given audience (Chau and Xu, 2012; Hsu et al., 2013). Bloggers can results (Statista, 2018), research in what we call now ‘influencer mar-
become opinion leaders and their posts can impact brands and products keting’ is still limited (De Veirman et al., 2017; Aral, 2011).
and influence potential customers (Smith et al., 2007). Therefore, we An important question arises: what makes a blogger, or any other

⁎
Corresponding author.
E-mail addresses: [email protected] (K. Sokolova), [email protected] (H. Kefi).

https://doi.org/10.1016/j.jretconser.2019.01.011
Received 15 September 2018; Received in revised form 15 January 2019; Accepted 15 January 2019
0969-6989/ © 2019 Elsevier Ltd. All rights reserved.

Please cite this article as: Karina Sokolova and Hajer Kefi, Journal of Retailing and Consumer Services,
https://doi.org/10.1016/j.jretconser.2019.01.011
K. Sokolova and H. Kefi Journal of Retailing and Consumer Services xxx (xxxx) xxxx

content creator, influence their followers? Businesses need to understand elaboration likelihood model of persuasion (ELM) (Petty and Cacioppo,
the ‘influencer’ phenomenon to make the right investment in influencer 1986), where the motivation and abilities of the person processing the
marketing and influencers. Also, the influencers need insights to be message would affect whether the argument quality or other peripheral
more persuasive while working with brands and products and pro- cues would influence the attitudes, intentions or beliefs of the receiver.
moting the products to their followers (Hollebeek, 2011; Hollebeek The information would be processed using a central route with cogni-
et al., 2014; Aral, 2011). tive responses related to information, facts and arguments or using a
Many researchers have focused on celebrity endorsers (Ford, 2018; peripheral route using peripheral cues: cognitive shortcuts that are easy
McCormick, 2016). But Social influencers are neither traditional ce- to process. Such cues could be related to the message and its pre-
lebrities, like famous actors or athletes, nor are their posts the tradi- sentation, to the source of the message or the speaker, the environment,
tional advertisements (Abidin, 2016). It was shown that bloggers with etc.
high audiences are more likable (De Veirman et al., 2017) and their The communicator and the way he/she is perceived can have an
‘promoting’ posts are better appreciated by potential customers than impact on the persuasion of an audience. Three characteristics of a
traditional ads (de Vries et al., 2012) even though the quality of the speaker are considered fundamental in persuasion: authority, cred-
content looks less professional (Hautz et al., 2014) or when the content ibility and social attractiveness, these are related to compliance, in-
is marked as ‘sponsored’ (Müller et al., 2018). Susarla et al. (2016) have ternalization and identification, respectively (Kelman, 1958).
noticed the ephemeral nature of attention and inequality in the popu- Authority is frequently related to the position of the communicator
larity of user-generated content. However, more knowledge and cues in the social structure. The communicator can have the ability to inflict
about the persuasiveness of YouTube and Instagram influencers are yet punishment or give a reward or approval to the message receiver who
to be discovered. complies in order to avoid punishment or to obtain a reward (Kelman,
The aim of this article is to fill in this gap by enhancing the com- 1958).
prehension of the persuasion cues related to social media influencers Credibility could be defined as the degree of the trustworthiness and
and their impacts on the attitudinal effects of influence attempts. We reliability of the source (Rogers and Bhowmik, 1970). Credibility is
focus on Instagram and YouTube bloggers of the beauty and fashion related to the perception of the quality of the communication by the
sector because of the intensive use of influencer marketing in this sector audience member. Credibility influences the audience through the in-
and within these two social network platforms (Garland and Reed, ternalization process: members share similar values and attitudes with
2018; FashionMonitor, 2016). the speaker (Kelman, 1958). Expertise - knowledge and experience the
Building on the persuasion theory developed by Petty and Cacioppo person has in the given domain - is one of the main factors of cred-
(1986), previous researchers found the importance of the arguments in ibility, along with trustworthiness, and goodwill (Hovland and Weiss,
the persuasion process, but, according to Mosler (2006), the peripheral 1951; McCroskey and Teven, 1999). Trustworthiness relates to the
cues could play a more important role than the arguments themselves in perceived honesty of the speaker and the goodwill reflects his/her
persuading a group. Users connect to social media websites to fulfill perceived caring about his/her audience. Generally, if the source is
their information, entertainment, and mood management needs (Shao, considered being trustworthy, attractive and perceived as an expert, it
2009). The content is entertaining and easy to absorb, adapted to the can influence the attitude and behavior of the audience (Ohanian,
young population that refers to the peripheral route of the Elaboration 1990; Petty and Wegener, 1998) including purchase intention
likelihood model (ELM) (Petty and Cacioppo, 1986). Although, the (Gunawan and Huarng, 2015).
physical attractiveness of a beauty and fashion blogger can also be On online social media, Lin et al. (2016) found that authority,
viewed as an argument for this particular industry. identity, and bandwagon cues explain the perception of credibility of
In this article, we focus on persuasion factors related to the beauty messages on Twitter. Credibility is also found to affect the content
and fashion influencers. We study how the credibility of the blogger, people choose to view, or to ignore, on social media, through what
para-social interaction, social and physical attractiveness and homo- Johnson and Kaye (2013) call selective exposure. Credibility were
phily attitude are related and affect the purchase intention. found to be related to the advertising value, and the advertisement
We base our study on an original survey-based dataset related to coming from the peer on social media is seen to be more credible
four highly followed beauty and fashion bloggers in France: comparing to traditional media Shareef et al. (2019). The audience in
EnjoyPhoenix, DazzlingDrew, Sananas2016 and Georgia Secret number of followers was found to be positively related to the credibility
(Horackova), who are all present and active on YouTube and Instagram. of the source on Twitter (Jin and Phua, 2014) and Instagram (De
The focus on women is motivated by the sector studied, in which this Veirman et al., 2017). The credibility of blogs was also found to affect
category of customers is a privileged target. Age differences will also be positively the attitude towards the blog (Colton, 2018).
investigated, as age has been recognized as a significant explanatory Social attractiveness refers to the likability of the speaker. Attractive
factor of the interaction mechanisms within social media and their speakers can change the attitudes of the audience (Sharma, 1990).
impacts in marketing (Goodrich, 2013; Muk, 2013). Regarding the same arguments used in a message, the attractive speaker
This paper is structured as follows. Section two presents the litera- is more likely to influence the audience than less appealing ones
ture review related to persuasion and para-social interaction theories. (Chaiken, 1979). Perceived social attractiveness can be a strong per-
Section three presents the research model and the hypotheses studied in ipheral cue for decision making.
this work. Section four describes the methodology used in this article: According to Kelman (1958), attractive speakers influence the au-
questionnaire development and administration as well as statistical dience through the identification process. An audience member would
procedure. Section five highlights the results of the global sample and feel similar or would like to be similar to the speaker and to create a
multi-group analyses. Section six proposes a discussion of the results positive relationship with him/her. Millenials - deemed to be extensive
and outlines the theoretical and practical implications of this work. online social network users - frequently identify with celebrities and
Section seven concludes the article and outlines limitations and also borrow some aspects of their personalities and lifestyle in order to look
future avenues to which this study opens the paths. like them (Peter, 2004; Boon, 2001). Thus, celebrities and online in-
fluencers launch fashion and other trends that are followed by those
2. Theoretical background who admire them. The impact of an action of a celebrity can be even
greater when the receivers perceive him/her as someone on whom they
2.1. Communicators and persuasion can personally rely (McCracken, 1989).
Similarity was also found to be a persuasion facilitating factor
While studying persuasion, researchers often refer to the (Brock, 1965), particularly if the similarity is relevant to the message

2
K. Sokolova and H. Kefi Journal of Retailing and Consumer Services xxx (xxxx) xxxx

(Berscheid, 1966). Similar people can inspire trust as the proposed according to (Rihl and Wegener, 2017). Hwang and Zhang (2018)
solutions could also work for the audience due to the similarities be- conducted a study in China and found that PSI is positively related to
tween the speaker and the audience member (Levine and Valle, 1975). purchase intention and influenced by empathy and the low self-esteem
Social cognitive theory suggests that people are more easily influenced of the follower.
by social figures that are perceived to be similar to them (Bandura,
1994). For instance, millenials are more likely to buy a product en- 3. Research model development
dorsed by someone matching their self-image, according to Sukhdial
et al. (2002). Although, the similarity could have an opposite effect if Building on our literature review, we have developed an original
the audience perceives a person as a non-expert (Goethals and Nelson, research model referring to the fundamental process of internalization
1973), particularly concerning factual matters. and identification applied to Instagram and YouTube influencers.
The similarity between interacting individuals in terms of beliefs, Internalization is related to the credibility of the speaker, while iden-
education and social status is called attitude homophily (Prisbell and tification is more an affection-related and emotional process that is
Andersen, 1980). This construct is related to the number of interactions strongly linked to para-social interaction (Kelman, 1958; Horton and
the individuals have, as similar communicators are more likely to in- Wohl, 1956; Brown and Basil, 2010).
teract with each other. The repeated interactions help to develop a We do not consider compliance in our model because influencers on
relationship or, in the case of a celebrity, para-social relationship that is social media are not direct authorities to their followers in the social
strongly related to the identification process (Brown and Basil, 2010). structure (Gass, 2015). Even if an influencer can be perceived as en-
dorsing a superior social status, a blogger cannot directly reward or
2.2. Influencers and para-social interaction punish the follower for purchasing a featured product or not doing so.
In our model, we aim to test the impact of credibility of the beauty
The theory of para-social interaction (PSI) defines the relationship influencer and para-social interaction the influencer is able to generate
between a spectator and a performer (Horton and Wohl, 1956) with an on the intention of their followers to purchase the product featured on
illusion of intimacy as for the ‘real’ interpersonal relationships (Dibble the channel. The follower could buy the product because the influencer
et al., 2015). Such a relationship is self-established and the other person is perceived to be reliable or believable as formulated by Wathen and
could be unaware of the relationship and influence it (Kelman, 1958). Burkell (2002). Following previous works (Fan and Miao, 2012;
Specifically, women having a para-social relationship with a ce- Sheeraz et al., 2012; Hui, 2017), we formulate our first research hy-
lebrity express the desire to have the same appearance and belong to pothesis:
the ‘club’ (Greenwood et al., 2008). In some ways, this could be similar H1: The influencer's credibility is positively related to the user
to the relationship created between social media users and online in- intention to purchase a featured product.
fluencers. Online social networking users can create such a relationship The follower might also feel a proximity with the influencer and
with bloggers by subscribing to their channels or blogs and following would buy a product because of this perceived proximity expressed and
their posts published on social media. Multiple followers can form an conceptualized, as we have seen in the previous section, in terms of
online community where the members share with the blogger similar para-social interaction which is reported in the literature as an ante-
values, beliefs and interests (Nambisan and Watt, 2011). YouTube cedent of the purchase intention (Kim et al., 2015; Lee and Watkins,
channels having one communicator are found to be more popular and 2016; Hwang and Zhang, 2018). Thus, our second research hypothesis
influential than channels having multiple speakers (Welbourne and is the following:
Grant, 2016), which could be due to para-social relationships created H2: Para-social interaction between the influencer and his/her
with the unique speaker. Celebrities that are able to connect with their followers is positively related to the user intention to purchase a
audience are more effective in terms of persuasion (Sukhdial et al., featured product.
2002) unlike unfamiliar celebrities (McCormick, 2016). Concerning the determinants of para-social interaction, Giles (2002)
The relationship between an influencer on Instagram, YouTube or has reminded that this concept is anchored in an asymmetric relation-
Facebook is not unidirectional, as users are able to add comments and ship between, what he calls media users (social networking sites' users)
discuss the content, and the content owner has the possibility of re- and media figures (such as social media influencers). This relationship
plying to the messages and comments related to that content. However, is built depending on a certain number of factors including how media
influencers cannot actually respond to all of their followers' requests users perceive media figures as attractive and similar to them. Attrac-
and engage fully in a true discussion due to a very high number of tiveness could be depicted as physical or social attractiveness (Rubin
followers and reactions they might have on their blogs. Accordingly, and McHugh, 1987; Rubin and Step, 2000).
online bloggers are similar to celebrities on traditional media. In the social psychology literature, physical attractiveness describes
Therefore, a fully bidirectional relationship and a ‘real’ friendship is how appealing or pleasing a person's physical features and aesthetic
more illusory than real between a blogger and a follower, as originally beauty are. These could vary across cultures and periods, but remain
put forward by the PSI definition in the context of traditional media strongly related to social interaction (Reis et al., 1980). In recent work
celebrities (Labrecque, 2014; Lee and Watkins, 2016). Kim et al. (2015) and in the context of influence online relationships, physical attrac-
found that social networking sites usage is positively related to the tiveness is reported as a predictor of PSI (Lee and Watkins, 2016).
development of a para-social relationship with celebrities. Therefore, we hypothesize:
For instance, young women follow both celebrities and bloggers on H3: Para-social interaction between the influencer and his/her
Instagram, both being influential on their followers. However, digital audience is positively to his/her physical attractiveness.
personalities seem to have stronger persuasiveness and credibility as Social attractiveness is defined as the likelihood to befriend or to
followers are more able to relate to them (Djafarova and Rushworth, choose the media figure as a social or a work partner. Along with
2017). The most appreciated profiles frequently support their followers physical attractiveness, social attractiveness has also been demon-
on Instagram, provide positive reviews and help them by reflecting a strated as a predictor of PSI concerning traditional (television, news-
positive image of themselves. Lee and Watkins (2016) demonstrated papers) or new media (social media) (Rubin and McHugh, 1987; Lee
that para-social interaction between YouTube bloggers and their fol- and Watkins, 2016; Kurtin et al., 2018). Thus, we define our fourth
lowers has a positive impact on luxury brand perception, and that hypothesis:
physical attractiveness and the perceived resemblance to the blogger is H4: Para-social interaction between the influencer and his/her
positively related to the para-social interaction. The feedback function audience is positively related to his/her social attractiveness.
available on YouTube does not influence the para-social interaction, Attitude homophily, defined as the degree to which people tend to

3
K. Sokolova and H. Kefi Journal of Retailing and Consumer Services xxx (xxxx) xxxx

bond with similar others, is deemed an explaining factor of inter- Table 1

personal communications and influence-based relationship patterns, Latent variables' reliability (CA=Cronbach's ; CR=Composite reliability;
especially in social networking studies (McPherson et al., 2001; Aral AVE=Average Variance Extracted).
and Walker, 2012; Walker, 2014). Based on previous works of (Giles, Latent variables CA CR AVE
2002; Eyal and Rubin, 2003; Lee and Watkins, 2016), it is also posited
as a determinant of PSI on social media. The next hypothesis we pro- Social Attractiveness 0,944 0,973 0,947
Physical Attractiveness 0,915 0,959 0,922
pose to test is:
Attitude Homophily 0.932 0.967 0.936
H5: Para-social interaction between the influencer and his/her Para-social Interaction 0,768 0,893 0,807
audience is positively related to his/her attitude homophily. Credibility 0,915 0,937 0,747
As for credibility, the literature describes it as a less subjective or Intention to Purchase 0,848 0,929 0,867
emotional-based determinant of social influence (Gass, 2015). It in-
cludes, according to Giffin (1967), expertness, reliability, intentions,
activeness and personal attractiveness. In our study related to the Table 2
Construct measures validity. Loading values are all significant at p < 0.01.
beauty and fashion industry, we propose to positively relate credibility
to physical attractiveness where it can serve as a selling point. For in- Latent variables with indicators Mean Sd Loading
stance, influencers can show, using their own appearance, how the
Social Attractiveness
results expected from the advocated products are confirmed. So that,
I think she could be a friend of mine (Q50) 3.42 1.133 0.974
the lack of physical appeal and fashion expertise could reduce the I would like to have a friendly chat with her (Q60) 3.52 1.123 0.972
perceived credibility of the influencer. We therefore hypothesize: Physical Attractiveness
H6: The credibility of the influencer is positively related to his/ I think she is quite pretty (Q70) 3.96 0.973 0.963
I find her very attractive physically (Q80) 3.72 0.972 0.957
her physical attractiveness.
Attitude Homophily
We consider the attitude homophily as a determinant factor of She thinks like me (Q90) 3.14 0.998 0.965
credibility as it refers to the perception of similarities between an in- She shares my values (Q100) 3.23 1.022 0.970
fluencer and a follower in terms of values. The internalization process Para-social Interaction
refers indeed to similar values that are found in another person. I look forward to watching her videos. reading her 3.20 1.238 0.934
posts and comments (Q200)
Therefore, we argue that attitude homophily would also be positively
I would follow her and interact with her on other social 2.74 1.295 0.861
related to the credibility of the influencer (Djafarova and Rushworth, networking sites (Q300)
2017). Credibility
H7: The credibility of the influencer is positively related to his/ I find this blogger expert in her domain (Q400) 3.48 1.036 0.869
I find this blogger efficient in her job (Q700) 3.08 1.108 0.848
her attitude homophily.
I find this blogger trustworthy (Q600) 3.48 1.064 0.916
Accordingly, our theoretical model includes six constructs and seven I think this blogger cares about her followers (Q500) 3.78 1.037 0.825
research hypotheses that are depicted in Fig. 1. It stipulates that cred- This blogger updates regularly her content (Q800) 3.54 1.054 0.871
ibility and PSI positively affect intention to purchase (respectively, Intention to Purchase
hypotheses H1, and H2); physical attractiveness, social attractiveness I would purchase the products promoted by the blogger 2.99 1.287 0.921
in the future (Q900)
and attitude homophily positively affect PSI (respectively, hypotheses
I would encourage people close to me to buy the 2.79 1.168 0.941
H3, H4 and H5). Whereas credibility is positively related to physical products promoted by the blogger (Q1000)
attractiveness (H6) and attitude homophily (H7).
As age has been revealed in the literature as a relevant variable to
differentiate influence patterns between generations (Hui, 2017; Merz operationalized using Lee and Watkins (2016), attitude homophily
et al., 2018), we also propose to test our model for different user age using Eyal and Rubin (2003), social and physical attractiveness with
categories, trying to assess the generational effect within the joint im- reference to Ohanian (1990). For the credibility construct, we adapted
pact of PSI and credibility on the purchase intention. Besides age, the the measurement of McCroskey and Teven (1999), by reducing the
second control variable used in our study is the influencer: we will test number of questions to improve the response rate of our study. Inten-
our model for the whole sample and also for each of the four influencers tion to purchase has been operationalized following See-To and Ho
studied. Gender is not considered as the whole population under study (2014). In conformance with the literature, all the items are reflexive
is female. and assessed with a 5-point Likert scale. Table 2 summarises the
questionnaire.
We collected data using an online survey administered within a
4. Methodology large population of users of social network platforms in France who are
interested in beauty and fashion content. To construct a more homo-
4.1. Questionnaire development and administration geneous sample we selected, using filtering questions, respondents who
are familiar with the following influencers: EnjoyPhoenix, Sananas,
To operationalize our constructs, all the measures used in this study GeorgiaSecret (a.k.a. Georgia Horackova) and Dazzlingdrew.
were adapted from the literature. Para-social interaction has been All of these four influencers are among the most popular beauty and
fashion influencers in France on YouTube and Instagram. Georgia
Horackova's blogs also cover fitness and vegan diet is addition to beauty
and fashion. All of these four influencers are females living in France
(between 20 and 35 years old): DazzlingDrew and EnjoyPhenix are
born in 1995; Sananas and Georgia Horackova are born in 1989. They
created YouTube blogs between 2010 and 2011 and have between 2
and 3 million (Sananas and EnjoyPhoenix) to 300–800 thousand fol-
lowers (Dazzlingdrew and Georgia Horackova) on YouTube. On
Instagram, EnjoyPhoenix, Sananas, GeorgiaSecret and Dazzlingdrew
count 4,3 million, 1,8 million, 289 thousand and 150 thousand fol-
lowers respectively. None of the four influencers are traditional
Fig. 1. Research model.

4
K. Sokolova and H. Kefi Journal of Retailing and Consumer Services xxx (xxxx) xxxx

celebrities or official experts on beauty and fashion. Table 4

79% of the sample is composed of users who are representative of Heterotrait-Monotrait ratio (HTMT) (Henseler et al., 2015).
the Z Generation (born after 1995), 17% of Y generation (born between 1 2 3 4 5
1980 and 1995), 4% belong to X Generation (born between 1960 and
1980). We have excluded from our study sample baby boomers born 1. Purchase Intention
2. Para-social Interaction 0,754
before 1960 (less than 1% of the original sample). There was no data
3. Physical Attractiveness 0,431 0,49
collected from the earlier generations in our study. 4. Social Attractiveness 0,58 0,76 0,705
A majority of respondents are female that is not surprising as beauty 5. Credibility 0,732 0,809 0,78 0,85
and fashion industry are mostly related to female field of interest. Less 6. Attitude homophily 0,648 0,805 0,636 0,842 0,833
than 1% of the responses were from male and these were excluded from
the study sample. The final sample included 1209 exclusively female
responses, distributed as follows: 304 for EnjoyPhoenix; 304 for Table 5
Sananas; 303 for Georgia Secret and 298 for Dazzlingdrew. Structural model evaluation (Global sample N = 1209). f 2 =effect size of path
coefficient; R2 =coefficient of determinants; Q 2 =predictive relevance/Stone-
Geisser Q 2 (blindfolding procedure with omission distance of 7).
4.2. Statistical procedures Construct/structural path f2 R2 Q2

A partial least squares (PLS) approach with SmartPLS software 3.2.7 Social Attractiveness
Social attractiveness Para-social interaction 0.079
(Ringle and Becker et al., 2015) was used to test the model. PLS is able
Physical Attractiveness
to model latent variables without requiring strict normal distributions Physical Attractiveness Para-social interaction 0.040
of data (Ringle et al., 2012). This approach is appropriate to this study Physical attractiveness Credibility 0.336
whose aim is the identification of the key drivers of constructs and not Attitude Homophily
theory confirmation that is best handled by covariance-based structural Attitude hompohily Para-social-interaction 0.160
Attitude homophily Credibility 0.634
modeling. We started with a confirmatory phase to assess the mea- Credibility 0.697 0.672
surement model. Credibility Purchase intention 0.164
We used factor analysis to verify the validity of all constructs. The Para-social interaction 0.520 0.450
results in Table 2 show the factor loading for all measures is above 0.8 Para-social interaction Purchase intention 0.118
Purchase intention 0.485 0.177
or above 0.9 allowing us to keep all measures. The internal consistency
of variables is verified using Cronbach's Alpha, the composite reliability
is also verified. The results are displayed in Table 1 showing high in-
the results.
ternal consistency. Average variance extracted (AVE) is higher that 0.5
that is adequate for convergent validity according to Hair et al. (2014).
To address discriminant validity, we first compare Average Variance 5. Results
Extracted (AVE) and Shared Variance between variables as suggested
by Fornell and Larcker (1981). The AVE of the latent constructs should First, we performed the analysis on the global sample. Table 5
be greater than the squared correlation between variables. Table 3 re- shows path coefficients and significance of the structural model for the
presents the related results where all of the square root of the AVE whole sample and Fig. 2 shows the results visually. With regard to the
(highlighted in bold) are greater than the correlations between variable levels of R2 , the model explains 48.5% of the variance of the intention
that confirms the discriminant validity of the constructs. We also ap- to purchase construct, 52% of the variance of PSI and 69.7% of the
plied heterotrait-monotrait ratio of correlations considered to be more variance of credibility, which is very satisfactory in social sciences
reliable to further confirm the discriminant validity (Henseler et al., (Chin, 1998).
2015). Table 4 shows satisfactory results where all values are below 0.9. Table 6 and Fig. 2 show the hypotheses testing results. All our hy-
We observe that some constructs, such as credibility, are correlated potheses are supported, except H5 (physical attractiveness para-so-
with other constructs that suggests it shares some of the predictive cial interaction) for which we have a significant path ( p value = 0.04 )
power over the dependent variables with them (Farrell, 2010). but it is unexpectedly negative rather than positive ( = 0.06 ), high-
Following the validation and reliability verification, we applied lighting that physical attractiveness is negatively related to para-social
bootstrapping analysis with 5000 re-samples to the whole sample to interaction. However, as expected, physical attractiveness is positively
examine the structural validity of the model (hypotheses testing). Then, related to the credibility of the blogger ( = 0.39).
we applied two multi-group analyses (MGA). First, we control the age Intention to purchase is determined by both para-social interaction
variable and investigate whether the generational effect is significant in and credibility where the impact of credibility is slightly higher than
this study. The second MGA is applied in order to test our hypotheses PSI ( = 0.4 and = 0.35 respectively). PSI is positively determined by
for each of the 4 influencers we are investigating, and then to compare attitude homophily and social attractiveness, where social

Table 3
Discriminant validity (Fornell and Larcker, 1981). The diagonal elements in
bold represent the square root of the average variance extracted for the re-
spective construct.
Correlations

1 2 3 4 5 6

1. Purchase Intention 0,931

2. Para-social Interaction 0,633 0,898
3. Physical Attractiveness 0,379 0,435 0,96
4. Social Attractiveness 0,521 0,666 0,655 0,973
5. Credibility 0,659 0,708 0,706 0,691 0,864
Fig. 2. Findings for the global sample (N = 1209). Accepted hypotheses are
6. Attitude homophily 0,579 0,694 0,587 0,69 0,673 0,968
highlighted in green.

5
K. Sokolova and H. Kefi Journal of Retailing and Consumer Services xxx (xxxx) xxxx

Table 6 relationship between social attractiveness and PSI for generation Z,

Results of hypothesis testing (Global sample N = 1209). One-tailed test. meaning that followers would not necessarily feel a friendship-like re-
Hypothesis/Structural path t-value p-value Result lationship with a blogger they regularly follow, as was supposed by
Berryman and Kavka (2017). That is not the case for the older gen-
H1 Credibility Purchase intention 0.409 14.202 0.000 Accepted erations (Y and X) where this path is significant. This could indicate a
H2 PSI Purchase intention 0.346 10.655 0.000 Accepted generational switch in the conception of para-social interaction, which
H3 Physical attractiveness PSI − 0.059 2.096 0.039 Rejected
H4 Social Attractiveness PSI 0.344 8.433 0.000 Accepted
becomes more like an addiction and content consumption than an il-
H5 Attitude homophily PSI 0.457 12.596 0.000 Accepted lusory relationship (Moghavvemi et al., 2017; Romo et al., 2017).
H6 Physical attractiveness 0.394 14.291 0.000 Accepted Shared values are certainly important for the youngest respondents
Credibility when they decide to regularly follow a blogger, unlike the attractive-
H7 Attitude homophily Credibility 0.541 20.623 0.000 Accepted
ness (social or physical) of the presenter. Even considering beauty and
fashion bloggers, the followers would not be more attracted to a more
beautiful blogger. This way, online social beauty bloggers are different
attractiveness has a slightly higher impact on PSI ( = 0.46 comparing
from traditional celebrities and luxury influencers, as the relationship
to = 0.34 ).
between physical attractiveness and PSI is not confirmed in the ma-
For further investigation, we performed multi-group analysis by
jority of tests in our study (Erdogan, 1999; Lee and Watkins, 2016).
respondent's age and by influencer; the results are displayed in Tables 7
However, multi-group analyses by influencer confirm that social
and 8 respectively. The results are globally consistent with the full
attractiveness is positively related to the para-social interaction for
sample analysis. H3 (physical attractiveness para-social interaction)
three out of four influencers: DazzlingDrew, EnjoyPhoenix and Sananas.
is rejected for all cases in both tests mostly because the statistical results
Georgia Secret differs from the other beauty influencers because her
are non significant: the null-hypothesis cannot be rejected. The path is
content focuses mostly on fitness and yoga, not covered by other three.
negative and significant only for one influencer Georgia Secretsyst
Georgia Horackova's YouTube channel provides multiple fitness tutor-
( = 0.18).
ials, and her Instagram feed features fitness-related clothes. She also
Analysis by generation shows a insignificant path for H4 (social
publishes fashion-related content like lookbooks and outfits, but does
attractiveness para-social interaction) for the generation Z (the
not publish makeup tutorials and rarely features makeup or beauty
youngest in our sample). The same hypothesis is not supported for
products. Additionally, she is the only blogger having a significant path
Georgia Secret.
between physical attractiveness and PSI. The relationship is negative
( = 0.184 ) suggesting the more the user finds her beautiful, the less
6. Discussion he/she is attached to her. The relationship between PSI and purchase
intention is stronger for her than for the other influencers ( = 0.635)
Both credibility and parasocial-interaction with bloggers impact the and the relationship between credibility and purchase intention is the
purchase intention of followers, meaning that a follower goes through weakest ( = 0.248). That is interesting, because she positions herself as
both affective and internalization processes when influenced to pur- a yoga instructor (suggesting she is an expert). Thus, her followers
chase a product. Global sample analysis highlights a slightly stronger probably look at Georgia's photos and videos for the content and not
impact of credibility on purchase intention. However, a closer look particularly for the speaker and they purchase the products when they
shows that the impact of para-social interaction is slightly stronger than are attracted to the content that could be pleasing or useful.
credibility for three out of four influencers and much stronger for the Unlike Lee and Watkins (2016), who found that physical attrac-
generation Z. It means that followers who are attached, or even ad- tiveness is related to PSI in the case of Youtube and luxury products, we
dicted, to the blogger are more likely to purchase than less attached found no, or a negative effect of physical attractiveness on PSI for
followers. This is particularly true for the younger generation: the beauty bloggers that are not particularly positioned as “luxury”. This
credibility of the blogger has a positive impact on purchase intention, counter-intuitive result concerning could, in fact, be explained by the
while having a much more silent role than PSI ( = 0.7 comparing to role played by bloggers and influencers on social media, which appears
the = 0.3 for the generation Z). Generation Y shows similar results, to be different from those played by models in traditional media, such
while the credibility becomes predominant for the older generation. as TV and magazines particularly in the luxury market. Models gen-
Homophily affects both para-social interaction and credibility, erally exhibit a perfect physical appearance and seldom ‘talk’ to people
showing that value sharing is a strong aspect of persuasion. Creating an directly, explain their preferences or provide useful day-to-day tips.
affective link with a blogger comes though shared values and those This is, on the contrary, exactly what bloggers do on social media. They
shared values are also indicative of the quality of the content the are therefore perceived as more authentic, as observed by Djafarova
blogger proposes. The shared values are related to the internalization and Rushworth (2017). This is why attitude homophily increases PSI as
process that seems to predominate in our results. expected, but not physical attractiveness. Thus, physical attractiveness
Interestingly, for younger generations, para-social interaction in the can be less important for customers seeking mass-market beauty and
case of beauty influencers and their followers is more like an addiction fashion products than for the luxury market where they are more likely
than a friendly relationship. We were not able to confirm a positive

Table 7
Results of hypothesis testing (Control Variable: Age). *** p < 0.001; ** p < 0.01; * p < 0.05. A=accepted; R=rejected. One-tailed test.

GenZ GenY GenX

Hypothesis t-value Result t-value Result t-value Result

H1 0.279 2.142* A 0.405 7.847*** A 0.405 7.847*** A

H2 0.697 6.715*** A 0.549 11.201*** A 0.309 8.240*** A
H3 −0.061 0.279 R −0.075 0.918 R −0.058 1.875 R
H4 0.030 0.134 R 0.289 2.496* A 0.356 7.806*** A
H5 0.518 2.344* A 0.515 4.793*** A 0.448 4.793*** A
H6 0.431 3.935*** A 0.396 5.850*** A 0.391 12.342*** A
H7 0.491 4.529*** A 0.566 4.532*** A 0.536 18.362*** A

6
K. Sokolova and H. Kefi Journal of Retailing and Consumer Services xxx (xxxx) xxxx

Table 8
Results of hypothesis testing (Control Variable: blogger). A=accepted; R=rejected. One-tailed test. *** p < 0.001; ** p < 0.01; * p < 0.05.

DazzlingDrew EnjoyPhoenix Georgia Secret Sananas

Hypothesis t-value Result t-value Result t-value Result t-value Result

H1 0.363 5.525*** A 0.338 8.228*** A 0.248 3.957*** A 0.395 9.353 * ** A

H2 0.223 3.373** A 0.545 12.371*** A 0.635 10.278*** A 0.532 13.118 * ** A
H3 0.031 0.523 R −0.051 0.770 R −0.184 2.974* R −0.060 1.064 R
H4 0.339 5.598*** A 0.228 3.251** A 0.194 1.457 R 0.239 2.638* A
H5 0.419 8.192*** A 0.474 7.514*** A 0.657 4.991*** A 0.371 5.617*** A
H6 0.263 5.221*** A 0.306 6.614*** A 0.329 4.375*** A 0.458 8.786*** A
H7 0.501 11.617*** A 0.543 11.858*** A 0.642 8.580*** A 0.509 12.433*** A

seeking higher quality. 6.2. Practical implications

Another explanation of this result could be in the complex and
subjective definition of what is ‘beautiful’. Beauty could be related to The results of this study could be beneficial for brands to help them
the body, the face, the hair or even the voice and how all these are establish affiliate relationships with influencers according to their
perceived (or magnified) on social media. Also, beauty bloggers can persuasive and para-social interaction talents. It is important for brands
wear different makeup, from natural to extravagant, that would be to understand the persuasive cues activated by the influences in order
perceived differently. For example, Sananas always wears strong ma- to define a fine balance between creativity and control. It is also ben-
keup, while DazzlingDrew and EnjoyPhenix mostly feature ‘natural’ and eficial to bloggers who wish to adopt more persuasive strategies while
less visible makeup. Georgia Secret opts for a makeup that is more in- creating the content and collaborating with brands.
tense than DazzlingDrew but more “every-day” than Sananas. A nega- Both credibility and PSI are positively related to the purchase in-
tive relationship between physical attractiveness and PSI for Georgia tention, meaning that the longtime followers of trustworthy influencers
can be explained by the fact that the makeup she always wears could be who care about their followers and show expertise on the subject are
considered beautiful but irrelevant to fitness. Also, some influencers more likely to purchase the featured products. For the younger audi-
show their faces without makeup, with natural imperfections that could ence, a connection making them look for further information posted by
make them seem “less beautiful” i.e. less perfect, but more “natural” their favorite blogger is more persuasive than credibility, confirming
and similar to normal people. the importance of creating such a link. Thus, a brand targeting young
Finally, the “beauty” could be related to the clothes a blogger wears population would find potential customers among the most attached or
and features in photos and videos. Followers would not necessarily like even addicted followers. As for bloggers, they should not only work on
all the outfits presented by the influencer, or consider them beautiful, community growth, but should also care about already acquired fol-
but could be generally attracted by the style of the person, which would lowers to build strong para-social relationships. Being socially attrac-
make them look for other posts and outfits. Then, the purchase inten- tive and friendly can help to reach generations Y and X. For the younger
tion could be affected by a similarity of a style and not by the beauty generation, the para-social relationship would more likely be related to
per see of the blogger (Kulmala et al., 2013). an ‘addictive’ content.
The perceived similarity with an influencer in terms of shared va-
lues is strongly related to both credibility and PSI. Brands should un-
6.1. Theoretical contribution derstand the values that are conveyed by the influencer and also valued
by the potential customers, as the bloggers would be more influential
Our study adds to knowledge on the recent research stream of in- on followers that think like them (for example, values related to the
fluencer marketing by investigating the combined effects of the influ- eco-behavior Dwivedi et al. (2015)). Therefore, the branded content
encers' perceived credibility and parasocial interaction on the purchase should be related to the same values. Influencers can use this insight
intention of users of social media platforms. It is one of the first studies and adopt attractive, common or trending values of the digital popu-
in this topic using a large data set (1209 respondents) collected from lation. Thus, trending values can be shared with a larger audience while
the audiences of four popular influencers in the beauty and fashion building both a para-social relationship with followers and enhancing
sector in France. To explain the online influence mechanisms, we have the credibility of the blogger. However, a significant change in values
applied theories originally developed for offline interactions, such as may have the opposite effect.
persuasion theory and para-social interaction theory, and discussed to The (perceived) expertise on the subject and the goodwill remain
what extent they help to enlighten the influential relationships between important for credibility and persuasion and should be transmitted via
an online content creator (who is not necessarily a deemed celebrity the created content, particularly while targeting the older generation.
outside social media context) and hos or her audience. While physical attractiveness is related to credibility of beauty and
This work contributes to the existing literature on the generational fashion bloggers, seeking a perfect physical appearance is not the ul-
effect on social media interactions (Goodrich, 2013; Muk, 2013), and timate goal of the followers.
how they have to be taken into consideration in influencer and social
media marketing. Our study pointed out, for example, that younger
generations value para-social interaction and what they perceive as 7. Conclusion
their personal attachment to the influencer more than the perceived
credibility. While this attachment is more related, for this category of Nowadays, companies are moving toward a more digitized com-
users, to homophily than it is to social or physical attractiveness. This munication to promote their products and brands, where social media
means that new canons of influence are experienced by younger gen- and influencer marketing become unavoidable when defining mar-
erations, especially generation Z, that need a more thorough ex- keting strategies.
amination, since many studies are currently focusing on Millennials. In this paper, we study how credibility and para-social interaction
are related to the purchase intention. Additionally, we study the in-
direct relationship of attitude homophily, physical attractiveness and
social attractiveness with the purchase intention. The results of this

7
K. Sokolova and H. Kefi Journal of Retailing and Consumer Services xxx (xxxx) xxxx

study could be beneficial for influencers helping them to adopt more processes for effective health interventions. Health Commun. 25 (6–7), 601–602
persuasive behavior. It is also important for brands to understand the (pMID: 20845160).
Chaiken, S., 1979. Communicator physical attractiveness and persuasion. J. Personal.
persuasive cues related to the influencers who present their products. Social. Psychol. 37 (8), 1387–1397.
From a theoretical point of view, our results suggest the need for Chang, Y.T., Yu, H., Lu, H.P., 2015. Persuasive messages, popularity cohesion, and
further studies on para-social interaction, that may need re-con- message diffusion in social media marketing. J. Bus. Res. 68 (4), 777–782.
Chau, M., Xu, J., 2012. Business intelligence in blogs: understanding consumer interac-
ceptualization regarding online influencers. We have not found any tions and communities. Mamagement Inf. Systels Q. 36 (4), 1189–1216.
evidence that physical attractiveness are related to PSI, particularly for Chin, W.W., 1998. The partial least squares approach to structural equation modeling. In:
young populations that seem not to value social attractiveness either, Marcoulides, G.A. (Ed.), Modern Methods for Business Research. Lawrence Erlbaum
Associates, Hillsdale, NJ, pp. 294–336.
probably meaning that the concept of PSI has changed. More research Colton, D.A., 2018. Antecedents of consumer attitudes' toward corporate blogs. J. Res.
could uncover what drives an audience to continue following the Interact. Mark. 12 (1), 94–104.
blogger and what attaches the younger generation to an influencer. De Veirman, M., Cauberghe, V., Hudders, L., 2017. Marketing through instagram influ-
encers: the impact of number of followers and product divergence on brand attitude.
Our study has some limitations related to its specific focus on a
Int. J. Advert. 36 (5), 798–828.
female audience located in France in a single sector: beauty and de Vries, L., Gensler, S., Leeflang, P.S., 2012. Popularity of brand posts on brand fan
fashion. This could restrict the generalization of the results to other pages: an investigation of the effects of social media marketing. J. Interact. Mark. 26
audiences and other sectors. Influencers we study in this work are (2), 83–91.
De Vries, N.J., Carlson, J., 2014. Examining the drivers and brand performance im-
present and popular on multiple platforms, such as YouTube and plications of customer engagement with brands in the social media environment.
Instagram, commonly posting different content. Thus, our respondents Dibble, J.L., Hartmann, T., Rosaen, S.F., 2015. Parasocial interaction and parasocial re-
could have been exposed to different types of content coming from the lationship: conceptual clarification and a critical assessment of measures. Hum.
Commun. Res. 42 (1), 21–44.
two platforms. Future work could investigate the persuasion clues for Djafarova, E., Rushworth, C., 2017. Exploring the credibility of online celebrities' in-
each platform separately considering their content specificity. Other stagram profiles in influencing the purchase decisions of young female users. Comput.
social media platforms specializing in more ephemeral interactions, Hum. Behav. 68, 1–7.
Dwivedi, Y.K., Kapoor, K.K., Chen, H., 2015. Social media marketing and advertising.
such as Snapchat, could be further studied to compare the results. Mark. Rev. 15 (3), 289–309.
It could also be interesting to further study the concept of beauty Dwivedi, Y.K., Kelly, G., Janssen, M., Rana, N.P., Slade, E.L., Clement, M., 2018. Social
and its different facets in the case of digital influence. Thus, not only media: the good, the bad, and the ugly. Inf. Syst. Front. 20 (3), 419–423.
Erdogan, B.Z., 1999. Celebrity endorsement: a literature review. J. Mark. Manag. 15 (4),
body and face, but voice, style, lifestyle, make-up or other more specific 291–314.
aspects of beauty could be relevant to the persuasion. Eyal, K., Rubin, A.M., 2003. Viewer aggression and homophily, identification, and
This study may also have been limited by the hypothetico-deductive parasocial relationships with television characters. J. Broadcast. Electron. Media 47
(1), 77–98.
research design adopted. We used a survey-based approach and adapted
Fan, Y.-W., Miao, Y.-F., 2012. Effect of electronic word-of-mouth on consumer purchase
items defined from the literature to measure psychometric variables intention: the perspective of gender differences. Int. J. Electron. Bus. Manag. 10 (3),
related to a complex and evolving psycho-sociological phenomenon. 175–181.
Future studies could address the subject using qualitative or mixed Farrell, A., 2010. Insufficient discriminant validity: a comment on bove, pervan, beatty,
and shiu (2009). J. Bus. Res. 63 (3), 324–327.
research designs to explore more emergent insights. Our study can be FashionMonitor, 2016. The Rise of Influencers - A Fashion and Beauty Monitor Report.
extended by a content analysis that could uncover arguments and FashionMonitor-Econsultancy, pp. 163–168.
content-related persuasion cues. The interaction between content gen- Ford, J.B., 2018. What do we know about celebrity endorsement in advertising? J. Advert.
Res. 58 (1), 1–2.
erated by an influencer and comments and reactions from the followers Fornell, C., Larcker, D.F., 1981. Evaluating structural equation models with unobservable
could be further investigated. variables and measurement error. J. Mark. Res. 18 (1), 39–50.
Influence encompasses dynamic and ubiquitous processes that Gallagher, K., 2018. The Influencer Marketing Report: research, strategy & platforms for
leveraging social media influencers. Bus. Inside. Fr.
continue to emerge and evolve, depending on the influencers, audi- Garland, C., Reed, M., 2018. How to measure the value of influencer marketing. Glob.
ences, brands and social media channels. More research is though Cosmet. Ind. 4(〈www.gcimagazine.com〉).
certainly needed to bring light on this new and complex phenomenon. Gass, R.H., 2015. Social Influence, Sociology of. In: International Encyclopedia of the
Social & Behavioral Sciences:Second Edition. pp. 348–354.
Gensler, S., Völckner, F., Liu-Thompkins, Y., Wiertz, C., 2013. Managing brands in the
Acknowledgements social media environment. J. Interact. Mark. 27 (4), 242–256.
Giffin, K., 1967. The contribution of studies of source credibility to a theory of inter-
personal trust in the communication process. Psychol. Bull. 68 (2), 104–120.
The authors would like to thank Marie Trassart, Alizee Ragot and
Giles, D.C., 2002. Parasocial interaction: a review of the literature and a model for future
Audrey Razafindrakoto for their help in collecting the data. research. Media Psychol. 4 (May), 279–305.
Goethals, G.R., Nelson, R.E., 1973. Similarity in the influence process: the belief-value
References distinction. J. Personal. Social. Psychol. 25 (1), 117–122.
Goodrich, K., 2013. Effects of age and time of day on Internet advertising outcomes. J.
Mark. Commun. 19 (4), 229–244.
Abidin, C., 2016. Visibility labour: engaging with influencers' fashion brands and #ootd Greenwood, D.N., Pietromonaco, P.R., Long, C.R., 2008. Young women's attachment style
advertorial campaigns on instagram. Media Int. Aust. 161 (1), 86–100. and interpersonal engagement with female tv stars. J. Social. Personal. Relatsh. 25
Alalwan, A.A., Rana, N.P., Dwivedi, Y.K., Algharabat, R., 2017. Social media in mar- (3), 387–407.
keting: A review and analysis of the existing literature. Gunawan, D.D., Huarng, K.-H., 2015. Viral effects of social network and media on con-
Aral, S., 2011. Commentary-identifying social influence: a comment on opinion leader- sumers' purchase intention. J. Bus. Res. 68 (11), 2237–2241.
ship and social contagion in new product diffusion. Mark. Sci. 30 (2), 217–223. Hair, J.F., Hult, Ringle, C., Sarstedt, M., 2014. A Primer on Partial Least Squares
Aral, S., Dellarocas, C., Godes, D., 2013. Introduction to the special issue-social media and Structural Equation Modeling (PLS-SEM). SAGE Publications (Incorporated).
business transformation: a framework for research. Inf. Syst. Res. 24 (1), 3–13. Hautz, J., Füller, J., Hutter, K., Thürridl, C., 2014. Let users generate your video ads? The
Aral, S., Walker, D., 2012. Supporting Online Material forIdentifying Inuential and impact of video source and quality on consumers' perceptions and intended beha-
Susceptible Members of SocialNetworks (June). viors. J. Interact. Mark. 28 (1), 1–15.
Bandura, A., 1994. Social cognitive theory of mass communication. In: Bryant, J., Hennig-Thurau, T., Gwinner, K.P., Walsh, G., Gremler, D.D., 2004. Electronic word-of-
Zillmann, D. (Eds.), LEA's Communication Series. Media Effects: Advances in Theory mouth via consumer-opinion platforms: What motivates consumers to articulate
and Research, pp. 61–90. themselves on the Internet?.
Berryman, R., Kavka, M., 2017. i guess a lot of people see me as a big sister or a friend: the Henseler, J., Ringle, C.M., Sarstedt, M., 2015. A new criterion for assessing discriminant
role of intimacy in the celebrification of beauty vloggers. J. Gend. Stud. 26 (3), validity in variance-based structural equation modeling. J. Acad. Mark. Sci. 43 (1),
307–320. 115–135.
Berscheid, E., 1966. Opinion change and communicator-communicatee similarity and Hollebeek, L., 2011. Exploring customer brand engagement: definition and themes. J.
dissimilarity. J. Personal. Social. Psychol. 4 (6), 670–680. Strateg. Mark. 19 (7), 555–573.
Boon, S., 2001. Admirer-celebrity relationships among young adults. Explaining percep- Hollebeek, L.D., Glynn, M.S., Brodie, R.J., 2014. Consumer brand engagement in social
tions of celebrity influence on identity. Hum. Commun. Res. 27 (3), 432–465. media: conceptualization, scale development and validation. J. Interact. Mark. 28 (2),
Brock, T.C., 1965. Communicator-recipient similarity and decision change. J. Personal. 149–165.
Social. Psychol. 1 (6), 650–654. Horton, D., Wohl, R.R., 1956. Mass communication and para-social interaction.
Brown, W.J., Basil, M.D., 2010. Parasocial interaction and identification: social change Psychiatry 19 (3), 215–229.

8
K. Sokolova and H. Kefi Journal of Retailing and Consumer Services xxx (xxxx) xxxx

Hovland, C.I., Weiss, W., 1951. The influence of source credibility on communication Peter, J.P., 2004. Consumer Behavior and Marketing Strategy. McGraw-Hill, Irwin/Series
effectiveness. Public Opin. Q. 15 (4), 635. in Marketing.
Hsu, C., Lin, J.C., Chiang, H., 2013. The effects of blogger recommendations on customers' Petty, R.E., Cacioppo, J.T., 1986. The elaboration likelihood model of persuasion. Vol. 19
online shopping intentions. Internet Res. 23 (1), 69–88. of Advances in Experimental Social Psychology. Academic Press, pp. 123 – 205.
Hui, T.X., 2017. The Effect of Source Credibility on Consumers' Purchase Intention in Petty, R.E., Wegener, D.T., 1998. Matching versus mismatching attitude functions: im-
Malaysia Online Community 1 (1), 12–20. plications for scrutiny of persuasive messages. Personal. Social. Psychol. Bull. 24 (3),
Hwang, K., Zhang, Q., 2018. Influence of parasocial relationship between digital celeb- 227–240.
rities and their followers on followers' purchase and electronic word-of-mouth in- Prisbell, M., Andersen, J.F., 1980. The importance of perceived homophily, level of un-
tentions, and persuasion knowledge. Comput. Hum. Behav. 87, 155–173. certainty, feeling good, safety, and self-disclosure in interpersonal relationships.
Jin, S.-A.A., Phua, J., 2014. Following celebrities' tweets about brands: the impact of Commun. Q. 28 (3), 22–33.
twitter-based electronic word-of-mouth on consumers' source credibility perception, Raggatt, M., Wright, C.J.C., Carrotte, E., Jenkinson, R., Mulgrew, K., Prichard, I., Lim,
buying intention, and social identification with celebrities. J. Advert. 43 (2), M.S.C., 2018. I aspire to look and feel healthy like the posts convey: engagement with
181–195. fitness inspiration on social media and perceptions of its influence on health and
Johnson, T.J., Kaye, B.K., 2013. The dark side of the boon? Credibility, selective exposure wellbeing. BMC Public Health 18 (1), 1002.
and the proliferation of online sources of political information. Comput. Hum. Behav. Reis, H.T., Nezlek, J., Wheeler, L., 1980. Physical attractiveness in social interaction. J.
29 (4), 1862–1871. Personal. Social. Psychol. 38 (4), 604–617.
Kapoor, K.K., Tamilmani, K., Rana, N.P., Patil, P., Dwivedi, Y.K., Nerur, S., 2018. Rihl, A., Wegener, C., 2017. Youtube celebrities and parasocial interaction: using feed-
Advances in social media research: past, present and future. Inf. Syst. Front. 20 (3), back channels in mediatized relationships. Convergence 0 (0) (1354856517736976).
531–558. Ringle, C.M., Sarstedt, M., Straub, D.W., 2012. Editor’s comments: a critical look at the
Kelman, H., 1958. Compliance, identification, and internalization: three processes of at- use of PLS-SEM in MIS quarterly. MIS Q. 36 (1) (iii–xiv).
titude change. J. Confl. Resolut. 2 (1), 51–60. Ringle, Christian M., W.S., Becker, J.-M., 2015. Smartpls 3. Bönningstedt: SmartPLS.
Kim, H., Ko, E., Kim, J., 2015. Sns users' para-social relationships with celebrities: social Retrieved from 〈http://www.smartpls.com〉.
media effects on purchase intentions. J. Glob. Sch. Mark. Sci. 25 (3), 279–294. Rogers, E.M., Bhowmik, D.K., 1970. Homophily-heterophily: relational concepts for
Klassen, K.M., Borleis, E.S., Brennan, L., Reid, M., McCaffrey, T.A., Lim, M.S., 2018. What communication research. Public Opin. Quat. 34 (4), 523–538.
people “like”: analysis of social media strategies used by food industry brands, life- Romo, L., Saleh, D., Coëffec, A., Chèze, N., Sedlackova-Morillon, H., Rigal, N., Taquet, P.,
style brands, and health promotion organizations on facebook and instagram. J. Med. Kern, L., Cottencin, O., 2017. Serials: Passion or Addiction? Evaluation Based on
Internet Res. 20 (6), e10227. Quantitative and Case Study. In: Serials and Addiction. Cambridge Scholar.
Knoll, J., 2016. Advertising in social media: a review of empirical evidence. Int. J. Advert. Rubin, A.M., Step, M.M., 2000. Impact of motivation, attraction, and parasocial inter-
35 (2), 266–300. action on talk radio listening. J. Broadcast. Electron. Media 44 (4), 635–654.
Kulmala, M., Mesiranta, N., Tuominen, P., 2013. Organic and amplified ewom in con- Rubin, R.B., McHugh, M.P., 1987. Development of parasocial interaction relationships. J.
sumer fashion blogs. J. Fash. Mark. Manag.: Int. J. 17 (1), 20–37. Broadcast. Electron. Media 31 (3), 279–292.
Kurtin, K.S., O'brien, N., Roy, D., Dam, L., 2018. The development of parasocial re- See-To, E.W.K., Ho, K.K.W., 2014. Value co-creation and purchase intention in social
lationships on YouTube. J. Social. Media Soc. Spring 7 (1), 233–252. network sites: the role of electronic word-of-mouth and trust - a theoretical analysis.
Labrecque, L.I., 2014. Fostering consumer-brand relationships in social media environ- Comput. Hum. Behav. 31 (February), 182–189.
ments: the role of parasocial interaction. J. Interact. Mark. 28 (2), 134–148. Shao, G., 2009. Understanding the appeal of user-generated media: a uses and gratifi-
Lee, J.E., Watkins, B., 2016. YouTube vloggers' influence on consumer luxury brand cation perspective. Internet Res. 19 (1), 7–25.
perceptions and intentions. J. Bus. Res. 69 (12), 5753–5760. Shareef, M.A., Mukerji, B., Dwivedi, Y.K., Rana, N.P., Islam, R., 2019. Social media
Levine, J.M., Valle, R.S., 1975. The convert as a credible communicator. Social. Behav. marketing: comparative effect of advertisement sources. J. Retail. Consum. Serv. 46,
Personal.: Int. J. 3 (1), 81–90. 58–69.
Lin, X., Spence, P.R., Lachlan, K.A., 2016. Social media and credibility indicators: the Sharma, A., 1990. The persuasive effect of salesperson credibility: conceptual and em-
effect of influence cues. Comput. Hum. Behav. 63, 264–271. pirical examination. J. Personal. Sell. Sales. Manag. 10 (4), 71–80.
Lipizzi, C., Iandoli, L., Ramirez Marquez, J.E., 2015. Extracting and evaluating con- Sheeraz, M., Iqbal, N., Ahmed, N., 2012. Impact of brand credibility and consumer values
versational patterns in social media: a socio-semantic analysis of customers' reactions on consumer purchase intentions in Pakistan. Int. J. Acad. Res. Bus. Social. Sci. 2 (8),
to the launch of new products using Twitter streams. Int. J. Inf. Manag. 35 (4), 1–10(URL 〈http://www.hrmars.com/admin/pics/961.pdf〉).
490–503. Shi, S., Chen, Y., Chow, W.S., 2016. Key values driving continued interaction on brand
McCormick, K., 2016. Celebrity endorsements: influence of a product-endorser match on pages in social media: an examination across genders. Comput. Hum. Behav. 62,
millennials attitudes and purchase intentions. J. Retail. Consum. Serv. 32, 39–45. 578–589.
McCracken, G., 1989. Who is the celebrity endorser? Cultural foundations of the en- Shiau, W.-L., Dwivedi, Y.K., Yang, H.S., 2017. Co-citation and cluster analyses of extant
dorsement process. J. Consum. Res. 16 (3), 310–321. literature on social networks. Int. J. Inf. Manag. 37 (5), 390–399.
McCroskey, J.C., Teven, J.J., 1999. Goodwill: a reexamination of the construct and its Smith, T., Coyle, J.R., Lightfoot, E., Scott, A., 2007. Reconsidering models of influence:
measurement. Commun. Monogr. 66 (1), 90–103. the relationship between consumer social networks and word-of-mouth effectiveness.
McPherson, M., Smith-Lovin, L., Cook, J.M., 2001. Birds of a feather: homophily in social J. Advert. Res. 47 (4), 387–397.
networks. Annu. Rev. Sociol. 27 (1), 415–444. Statista, 2018. Statista dossier on influencer marketing in the United States and world-
Merz, M.A., Zarantonello, L., Grappi, S., 2018. How valuable are your customers in the wide.
brand value co-creation process? The development of a Customer Co-Creation Value Sukhdial, A., Aiken, D., Kahle, L., 2002. Are you old school? J. Advert. Res. 42 (4), 71–81.
(CCCV) scale. J. Bus. Res. 82, 79–89 (September 2017). Susarla, A., Oh, J.H., Tan, Y., 2016. Influentials, imitables, or susceptibles? Virality and
Moghavvemi, S., Sulaiman, A.B., Jaafar, N.I.B., Kasem, N., July 2017. Facebook and word-of-mouth conversations in online social networks. J. Manag. Inf. Syst. 33 (1),
youtube addiction: The usage pattern of malaysian students. In: 2017 International 139–170.
Conference on Research and Innovation in Information Systems (ICRIIS).pp. 1–6. Uzunoğlu, E., Kip, S.M., 2014. Brand communication through digital influencers: lever-
Mosler, H.-J., 2006. Better be convincing or better be stylish? A theory based multi-agent aging blogger engagement. Int. J. Inf. Manag. 34 (5), 592–602.
simulation to explain minority influence in groups via arguments or via peripheral Walker, D., 2014. Tie Strength, Embeddedness, and SocialInuence: A Large-Scale
cues. J. Artif. Soc. Social. Simul. 9 (3), 4. Networked Experiment. pp. 1–19.
Muk, A., 2013. What factors influence millennials to like brand pages? J. Mark. Anal. 1, Wathen, C.N., Burkell, J., 2002. Believe it or not: factors influencing credibility on the
127–137. Web. J. Am. Soc. Inf. Sci. Technol. 53 (2), 134–144.
Müller, L.K., Mattke, J., Maier, C., 2018. #sponsored #ad: exploring the effect of influ- Welbourne, D.J., Grant, W.J., 2016. Science communication on youtube: factors that
encer marketing on purchase intention. AMCIS. Assoc. Inf. Syst. affect channel and video popularity. Public Underst. Sci. 25 (6), 706–718 (pMID:
Nambisan, P., Watt, J.H., 2011. Managing customer experiences in online product com- 25698225).
munities. J. Bus. Res. 64 (8), 889–895. Zhu, Z.b., Su, J., Kong, L., 2015. Measuring influence in online social network based on
Ohanian, R., 1990. Construction and validation of a scale to measure celebrity endorsers' the user-content bipartite graph. Comput. Hum. Behav. 52, 184–189.
perceived expertise, trustworthiness, and attractiveness. J. Advert. 19 (3), 39–52.