Journal of Business Research xxx (xxxx) xxx–xxx

Contents lists available at ScienceDirect

Journal of Business Research

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

Understanding consumers’ social media engagement behaviour: An

examination of the moderation effect of social media context
Dongmei Caoa, , Maureen Meadowsa, Donna Wongb, Senmao Xiaa
⁎

a
Faculty of Business & Law, Coventry University, United Kingdom
b
Graduate School of Sport Sciences, Waseda University, Tokyo, Japan

1. Introduction low-value information including spam and fake news (Fulgoni &
Lipsman, 2017; Mintel, 2018a; BBC, 2018). Thus, one may wonder how
Consumers' Social Media Engagement Behaviour (SMEB) is defined consumers' trust in - or scepticism towards - SMM content affects their
in this study as consumer engagement behaviour with social media engagement behaviour on social media sites. This is a fundamental
marketing content, including brand-generated and peer-generated issue to be addressed by digital marketers seeking to encourage active
content, on social media platforms such as Facebook and YouTube consumer engagement. Understanding consumer engagement is critical
(Popovic, 2016). Previous studies have explored engagement with for digital marketing companies in order to build and maintain cus-
content generated by luxury fashion brands on social media platforms, tomer loyalty and to attract new customers. However, there is a lack of
where consumers 'Follow', 'Like', 'Comment' on or 'Share' content empirical research on the effect of potentially important contextual
(Kumar, Bezawada, Rishika, Janakiraman, & Kannan, 2016; Pentina, factors, such as trustworthiness, associated with SMM. The richness of a
Guilloux, & Micu, 2018). Social media users interact and network with medium is another such contextual factor, which must be considered in
peers in a brand social media community by 'Commenting' on peers’ order to understand consumers' SMEB with different social networking
'Posts' or responding to peers’ 'Reviews' of their shopping experience sites. Studies show that the depth of consumer engagement varies on
(Schau, Muñiz, & Arnould, 2009; Habibi, Laroche, & Richard, 2016). various social media such as Instagram and Twitter (Guidry, Jin, Orr,
Consumers' SMEB within a social media community enables them to Messner, & Meganck, 2017), Facebook (Wang, Kim, Xiao, & Jung,
share information and build up their social networks (Dessart, 2017), and YouTube (Khan, 2017). For instance, Instagram, due to its
Veloutsou, & Morgan-Thomas, 2015; Khan, 2017). unique attributes and functionality associated with high-quality visual
In the digital era, the use of social media platforms such as content, appears to engage users more intensely, compared to other
Facebook, Twitter, Instagram, Pinterest, and YouTube has become a social networking sites such as Facebook and Twitter (Mintel, 2018a;
norm for fashion retailers seeking to engage consumers actively. In their Pentina et al., 2018). Therefore, digital marketers require a fuller un-
Social Media Marketing (SMM), retailers provide live updates via derstanding of how the attributes of social media, such as media rich-
newsfeed, vlogs, tweets, photos and posts on arrivals of new products or ness and trustworthiness, can create variations in consumer engage-
services (Pentina et al., 2018). From consumer engagement through ment behaviour.
social media, fashion retailers gain brand association and increased This research aims to understand the effects of media richness and
brand performance (Rapp, Beitelspacher, Grewal, & Hughes, 2013; content trustworthiness on the SMEB of consumers. To this end, we
Fulgoni, 2015; Habibi et al., 2016; Dessart, 2017), influencing purchase develop our research model by referring to the Theory of Planned
decisions and sales (Muralidharan & Men, 2015; Kumar et al., 2016; Behaviour (TPB) (Ajzen & Fishbein, 1980; Ajzen, 1991; Venkatesh,
Pentina et al., 2018). Meanwhile, consumers are motivated to engage Morris, Davis, & Davis, 2003). It is important to point out that this study
with SMM within a social media community to obtain and share in- approaches SMEB from a social media contextual perspective, in order
formation, to learn about the latest fashion trends and to keep them- to provide fashion brands and social media marketers with a better
selves informed about new products from a favourite brand and event understanding of how media richness and content trustworthiness in-
updates (Dessart, 2017; Mintel, 2018b). fluences the SMEB of consumers. Media richness and content trust-
While many consumers are motivated to interact with peer users on worthiness are viewed as contextual factors in the TPB model, moder-
social media sites (Wang & Yu, 2017), there is also growing evidence ating the relationship between engagement intention and engagement
that some consumers have chosen not to trust or engage with SMM due behaviour. A detailed discussion of the research model is presented in
to various forms of 'digital pollution', i.e. an overload of redundant, section 2.4.

Corresponding author at: William Morris Building (reception), Coventry University, Coventry CV1 5DL, United Kingdom.
⁎

E-mail addresses: [email protected] (D. Cao), [email protected] (M. Meadows), [email protected] (D. Wong),
[email protected] (S. Xia).

https://doi.org/10.1016/j.jbusres.2020.06.025
Received 9 September 2019; Received in revised form 9 June 2020; Accepted 11 June 2020
0148-2963/ © 2020 Published by Elsevier Inc.

Please cite this article as: Dongmei Cao, et al., Journal of Business Research, https://doi.org/10.1016/j.jbusres.2020.06.025
D. Cao, et al. Journal of Business Research xxx (xxxx) xxx–xxx

Fashion social media marketing is an appropriate context for this the mental processes involved in focusing on intense attention and
study, as it aims to explain consumers' SMEB from the social media absorption. Affective engagement connotes emotional reactions, such as
contextual perspective, which has been the subject of little research but enthusiasm and enjoyment. Behavioural engagement is defined as the
of practical importance (Dahlan, 2005). The fashion industry is highly active manifestations of the engagement concept, which include
dynamic, and consumers' shopping behaviour is constantly changing. sharing, learning and endorsing behaviours. This behavioural perspec-
To react to the changing market, fashion retailers constantly seek new tive is thought to be more practical and in line with the analytics me-
ways to rapidly understand emerging trends and engage with con- trics used to measure SMM performance; it also offers more actionable
sumers (Bendoni, 2017). In an era of digital transformation, social insights (Pentina et al., 2018), which this study seeks to follow and
media networking provides a new approach for fashion brands and extend.
fashion retailers to interact with consumers actively; it also represents a In an early attempt, Muntinga et al. (2011) introduced three types of
new trend in marketing and brand management practices for many brand-related social media usage, namely consuming, contributing, and
fashion brands (Kim & Ko, 2012; Phan & Park, 2014). creating. Schivinski, Christodoulides, and Dabrowski (2016) developed
This study makes theoretical contributions and offers practical im- and empirically examined the same three-factor framework measuring
plications regarding social media engagement with fashion SMM. First, consumers’ engagement activities with brand-related social media
we contribute to the current understanding of consumers' social media content. Taking this scale further, Pentina et al. (2018) conducted a
engagement as influenced by social media contextual factors, measured study that suggests that various consumer-brand engagement beha-
by media richness and content trustworthiness. Such contextual factors viours in social media sites represent different levels of engagement
have remained underexplored in the existing literature, despite some effort and creativity, e.g., from “following” (lower level) to “com-
investigation of antecedents of SMEB from other perspectives such as menting” (higher level). 'Consumption' represents a minimum level of
psychology (Hwong, Oliver, Van Kranendonk, Sammut, & Seroussi, engagement, in which users passively consume SMM content, e.g.
2017), personality (Lim, Hwang, Kim, & Biocca, 2015), society (Lim reading a fashion blog or watching a video posted by a fashion brand.
et al., 2015), culture (Bail, 2016; Alt, 2017a), or a combination of these Contribution entails a higher level of engagement, involving peer-to-
dimensions (see for instance Dessart, 2017; Zhang, Borden, & Kim, peer and peer-to-content interactions on social networking sites, e.g.
2018; Alt, 2017a). Second, there is a lack of empirical studies focusing commenting on a post or forwarding it to peers. This level of engage-
on variations in SMEB. Drawing on the extant literature, we first con- ment leads to the wider dissemination of social media content, which
ceptualise SMEB and then operationalise and empirically examine could be retailer-generated or peer consumer-generated. Creation is the
variations in SMEB by referring to TPB. Third, we contribute to the highest level of engagement, in which consumer-generated content is
literature on TPB, which has thus far been applied to explain and pre- published on social networking sites, e.g. consumers create and post a
dict human behaviour and behavioural intention (e.g., Cooke & French, picture, vlog or publish a fashion review or opinion piece. This research
2008; Al-Debei, Al-Lozi, & Papazafeiropoulou, 2013). Empirical re- considers all three levels of engagement behaviours - consumption,
search has mostly investigated the antecedents of behavioural intention contribution, and creation.
or actual behaviour; little research has investigated variations in the The TPB logic (Cooke & French, 2008), supported by extant em-
relationship between behavioural intention and actual behaviour. Also, pirical studies (e.g., Muralidharan & Men, 2015; Oliveira, Huertas, &
studies have typically focused on one single behavioural response while Lin, 2016), suggests that a specific behavioural intention is positively
there could be several (Ajzen & Sheikh, 2013). Our research seeks to related to the corresponding actual behaviour. However, whether the
extend knowledge on TPB by addressing these two gaps. Therefore, we specific behavioural intention can explain and predict a number of re-
contribute to the understanding of social media engagement associated levant behaviours have rarely been examined in extant literature. In
with digital media marketing, and also to the advancement of the TPB this research context, fashion consumers are likely to draw inspiration
literature. Besides, the findings of our research will help fashion brands from online sources such as a brand's official websites or official ac-
and social media marketers to manage consumer engagement with counts on social media sites. Engagement with fashion retailers and
SMM more effectively. For instance, when selecting a social media peers in a fashion brand's social media communities allows fashion
platform, digital fashion marketers should consider the fit between consumers to share and communicate about fashion trends and new
their specific digital marketing purpose and the richness of the different products. This research thus examines whether fashion consumers' en-
media chosen. Our results also have implications for digital fashion gagement intention with fashion SMM is enacted to perform the three
marketers with regard to the content and trustworthiness of the mate- levels of engagement behaviours – consumption, contribution, and crea-
rial used to engage social media consumers. tion. We posit the following three hypotheses:
The remainder of this study is structured as follows. In the next
H1a. Engagement intention is positively related to consumption
section, we develop hypotheses and formulate our research model. This
behaviour.
is followed by the introduction of our research methods. Next, we re-
port and discuss our findings. Finally, we conclude with the theoretical H1b. Proof Engagement intention is positively related to contribution
contributions and practical implications. behaviour.
H1c. Proof Engagement intention is positively related to creation
2. Research hypotheses and model building
behaviour.
2.1. Social media engagement behaviours
2.2. Moderation effects of media richness
SMEB comprises two key elements, namely, social media and con-
sumers' engagement behaviour. Social media is defined as a group of The concept of media richness is rooted in the media richness theory
Internet-based applications that allow the creation and exchange of (Brunelle, 2009). Media richness theory (MRT) is a widely cited in-
user-generated content (Kaplan & Haenlein, 2010). Numerous types of formation processing theory that explains media usage and commu-
social media fall within this broad definition such as Facebook, In- nication effectiveness. Introduced by Daft and Lengel (1986), the MRT
stagram, Pinterest, Twitter, Google Plus, WhatsApp, and YouTube. proposes a hierarchy of information media based on media richness.
Scholars (Muntinga, Moorman, & Smit, 2011; Dessart et al., 2015; The four factors to distinguish the richness level of a medium are the
Dessart, 2017) are generally in agreement that three dimensions - feedback capability of the medium; the number of channels used such
cognitive, affective and behavioural - constitute consumer engagement. as email and face-to-face communication; the source of information -
Dessart (2017) conceptualises cognitive engagement as being akin to personal (e.g., relatives and friends) or impersonal (e.g., retailers); and

2
D. Cao, et al. Journal of Business Research xxx (xxxx) xxx–xxx

finally, language variety such as verbal or non-verbal (e.g., body lan- are revealed between the intentions and actual behaviours of con-
guage and photos). MRT evaluates the ability of a medium to ade- sumers, lead to calls for further research (e.g. Cao, Tao, Wang, Tarhini,
quately communicate a complex message (Carlson & Zmud, 1999). The & Xia, 2019). Ajzen and Sheikh (2013) suggest that variations between
selection of a medium to effectively convey a message is mainly de- behavioural intention and actual behaviour might be related to alter-
termined by the characteristics of the message and the audience, e.g., native behaviours rather than a single act, while Wang, Zhang, Yu, and
complex vs simple, personal vs impersonal, single vs large audience Hu (2018) suggest that such differences might be due to the moderation
(Daft & Lengel, 1986). Extant MRT literature has focused pre- effect of situational or conditional variables between intention and
dominantly on media choices rather than media engagement behaviour behaviour. Wang et al. (2018) also argue that, in the case of the re-
(Kahai & Cooper, 2003). This research seeks to extend knowledge in the sponsible environmental behaviour (REB) of tourists, past studies have
latter area. MRT originates from research based on traditional com- neglected the influence of situational factors on the relationship be-
munication media such as emails and phone calls rather than inter- tween REB intention and actual behaviour. In their study, the moder-
active web-based social media. There is a lack of research that eluci- ating role of environmental interpretations was verified.
dates the meaning of ‘richness’ in the current context of digital In this study, we contend that the richness levels of the social media
communication and engagement. sites adopted by fashion brands and fashion marketers can play an
According to the Mintel database (Mintel, 2018b), Facebook, important role in the translation of the engagement intention of fashion
Twitter, YouTube, Instagram and Pinterest are the most popular social consumers and social media users into various levels of engagement
media networks used by marketers to target consumers of online behaviours in the fashion brand virtual community, that is, consump-
shopping in the UK. From the media richness point of view in a fashion tion, contribution, and creation activities associated with fashion SMM.
context, these five platforms possess varied characteristics of richness in Our statement is in line with the assertion of media richness theory that
terms of their feedback capability (e.g., the immediacy of feedback), task performance will be improved when a task matches the richness of
multiple information cues (e.g., text, video and audio), and available the chosen medium (Daft & Lengel, 1986). However, little is known
communication tools (e.g. like, share, comment, post). Among the five, about the moderation effect of the richness of social media on the re-
Facebook is often considered the richest platform since it encompasses lationship between fashion consumers' engagement intention and their
all of the above capabilities; Facebook is viewed as easy-to-use and corresponding engagement activities with a fashion social media com-
engaging (Wang et al., 2017). Twitter is deemed to be less ‘rich’ due to munity. Therefore, we formulate the following three hypotheses:
its character limit. Information contained in tweets can also be quickly
H3a. Engagement intention is more likely to be related to consumption
and easily ‘drowned out’ by new incoming messages. YouTube is a
in a positive direction when media are richer than when media are
video network that does not feature highly regarding community in-
leaner.
teractivity and feedback. Instagram and Pinterest are gaining in popu-
larity due to their high-quality visual information for SMM; however, H3b. Engagement intention is more likely to be related to contribution
they offer fewer opportunities for instant feedback or text comment in in a positive direction when media are richer than when media are
comparison to other social media platforms (Kim, Seely, & Jung, 2017). leaner.
The richness of social networking sites becomes increasingly im-
H3c. Engagement intention is more likely to be related to creation in a
portant in the fashion context as it can foster social and interactive
positive direction when media are richer than when media are leaner.
experiences beyond core purchase behaviour (Malthouse & Hofacker,
2010; Hollebeek, Glynn, & Brodie, 2014). The varying richness of the
fashion social media platforms allows fashion consumers to use an array 2.3. Moderation effects of content trustworthiness
of tools and resources on social media to engage with fashion brands,
resulting in different levels of engagement. For example, consumers can Content trustworthiness draws upon the source credibility theory
choose to consume fashion brand-related media just by viewing the (Chang, Yu, & Lu, 2015). Hovland and Weiss (1951) found that people
marketed content on Facebook; by further 'Commenting' on a fashion tend to discount information from untrustworthy sources; the dis-
post or ‘Liking’ the post, they are moving from the role of a fashion association of content from its untrustworthy source is likely to result in
consumer to a contributor to the social media content in the brand higher acceptance of untrustworthy information. The source credibility
community. As Malthouse and Hofacker (2010) suggest, the richness of theory has been applied in marketing and communication research,
social media enables interactive marketing, such as value co-creation, where studies have examined the effect of trustworthiness as a key
which encourages long-term consumer engagement with fashion source of credibility on various behavioural intentions and attitudes
brands. The creation dimension represents the strongest level of online (e.g., Gefen, 2002; Kim & Kim, 2014; Ayeh, 2015). For example, peo-
brand-related engagement, and the most likely to build up brand loy- ple's attitudes toward a target issue became more favourable when they
alty (Muntinga et al., 2011). At the same time, the richness of media received a message from a source with higher credibility (Tormala &
enables fashion consumers' engagement within a virtual social setting, Clarkson, 2007).
where fashion consumers can interact with other brand community Source credibility has become increasingly pertinent in the new
members (Brodie, Ilic, Juric, & Holleneek, 2011; Kuo & Feng, 2013; media environment where information can be created and disseminated
Dessart et al., 2015; Habibi et al., 2016). From this discussion, the by amateurs and unverified sources. The freedom to publish in the
following three hypotheses are formulated: public domain with no quality control and validation process, aided by
the ease of and access to publishing, have affected the credibility of
H2a. Media richness is positively related to consumption behaviour.
information disseminated and received. This has attracted increasing
H2b. Media richness is positively related to contribution behaviour. levels of attention from scholars interested in the source credibility of
various social media (Braten, Braasch, Strømsø, & Ferguson, 2015).
H2c. Media richness is positively related to creation behaviour.
Research suggests that retail-generated marketing content, such as ad-
From the empirical TPB literature, the extent to which an actual vertisements published on social media sites (e.g. Facebook and
behaviour is determined by the corresponding behavioural intention Twitter), is less credible than content on traditional media (e.g. TV and
seems to vary significantly (Ajzen & Sheikh, 2013). A meta-analysis of radio) (Prendergast, Liu, & Poon, 2009; Lai & Liu, 2020). A recent study
the empirical literature on TPB by Sandberg and Conner (2008) sug- of social media security and trustworthiness has suggested that trust
gests that, on average, 54 per cent of intentions are predicted by ex- between users is of vital importance for the establishment of a credible
isting models. In contrast, intentions can explain only 41 per cent of social networking site (Zhang & Gupta, 2018). However, there is a need
actual behaviours. These differences, and the potential paradoxes that to examine whether the trustworthiness of media content embedded on

3
D. Cao, et al. Journal of Business Research xxx (xxxx) xxx–xxx

Social media context

• Media richness
• Content trustworthiness

Engagement behaviour
Engagement • Consumption
intention • Contribution
• Creation

Fig. 1. Conceptual model.

social media affects consumer engagement behaviour (Swani, Milne, H5c. Engagement intention is more likely to be related to creation in a
Brown, Assaf, & Donthu, 2017). positive direction when content is more trustworthy than when content
In the instance of fashion retailing, the trustworthiness of fashion is less trustworthy.
SMM content has been called into question partly because of re-
presentation issues. It is difficult for consumers to gain complete and
accurate information, such as the exact measurement or colour of a 2.4. Research model
fashion item, or the 'feel' of the fabric, since they have to rely on images
posted online by retailers, or peer-consumers' subjective comments In the field of social media and digital marketing, many analytical
(Fulgoni & Lipsman, 2017). Also, consumer-generated content is tools and techniques are increasing in popularity, such as neural net-
sometimes viewed with scepticism; peers are not considered to be a works, big data analytics, trend analysis, and network analysis, many of
reliable source of information due to their perceived lack of expertise, which have contributed to exploration and visualisation of the domain
according to the source credibility theory (Ayeh, 2015). Therefore, we phenomena (Richard, 2015). However, traditional SEM models and
expect that the trustworthiness of SMM content will positively influence analytical techniques are still dominant in social and behavioural re-
all three levels of engagement behaviours. Hence, we propose the fol- search due to their maturity and strong explanatory power in predicting
lowing three hypotheses: human behaviour (Teng, Khong, & Goh, 2015).
TPB offers an appropriate reference point for the theoretical un-
H4a. Content trustworthiness is positively related to consumption derpinnings and development of our model. TPB seeks explanations
behaviour. from the actors' cognitive, social, and psychological characteristics for a
H4b. Content trustworthiness is positively related to contribution specific behavioural intention, which subsequently leads to the corre-
behaviour. sponding actual behaviour (Ajzen & Fishbein, 1980; Ajzen, 1991). Prior
studies have provided valuable insights about explanatory factors for a
H4c. Content trustworthiness is positively related to creation behaviour. certain behavioural intention, such as adoption and use of information
Following on with the exploration of consumer engagement inten- technology (Venkatesh & Bala, 2008). However, such studies have been
tion, we seek an explanation from the contextual factors on the varia- criticised due to the weak predictive power of relationships suggested
tions in the relationship between the engagement intention and actual between behavioural intention and actual behaviour (Teng et al.,
behaviour as advocated by Ajzen and Sheikh (2013) and Wang et al. 2015).
(2018). In addition to media richness, we propose content trust- Drawing on the existing literature, we develop a conceptual model
worthiness as another contextual cue. The existing literature suggests for this study, as shown in Fig. 1. The dependent variable - engagement
that little is known about the mediating role played by the content or behaviour - is composed of three types of engagement behaviours,
content characteristics of social media. In an early experimental study, namely consumption, contribution and creation (Schivinski et al., 2016),
Dahlen (2005) examined media as a contextual cue, and the results in association with SMM (Fig. 1). Here, we examine how the social
suggest that the creative media choice enhances perceived brand as- media context influences the relationship between consumers' engage-
sociations and increases the credibility of advertising and positive ment intention and their realised engagement behaviour. Bronner and
brand attitudes. In a recent study, Khobzi, Lau, and Cheung (2019) Neijens (2006) suggest that the social media context could include at-
examined the moderation effect of users' ‘thumbs-up’ and ‘reply’ on the tributes of a social medium, as well as the content embedded in the
relationship between message (i.e. advertisement post) framing and medium. Media richness measuring characteristics of social media, and
users' online social interactions within a social media brand commu- content trustworthiness measuring media content, are thus utilised as
nity. The results suggested that when neutrally framed comments are proxies of social media context, where these two factors are theoreti-
left on a brand-generated post, this is not likely to help the post to cally underpinned within the domain of Information and Communica-
obtain a higher level of engagement. Little is known about the mod- tion Technologies (ICT).
eration effect of the content trustworthiness of social media on the re-
lationship between fashion consumers' engagement intention and their 3. Methodology
corresponding engagement activities on a fashion social media com-
munity. Hence, we put forward the following three hypotheses: 3.1. Measures

H5a. Engagement intention is more likely to be related to consumption

The measures of the constructs were drawn primarily from extant
in a positive direction when content is more trustworthy than when
literature and adapted to the research context of fashion SMM. Five-
content is less trustworthy.
point Likert or semantically differential scales were used for measure-
H5b. Engagement intention is more likely to be related to contribution ment. Descriptive statistics in terms of means and standard deviations
in a positive direction when content is more trustworthy than when of the measures are displayed in Table 1.
content is less trustworthy. For the four items measuring engagement intention, three out of the
four were adapted from Taylor and Todd (1995). Paying attention to

4
D. Cao, et al. Journal of Business Research xxx (xxxx) xxx–xxx

Table 1 Table 2
Constructs and measures. Sample demographics (n = 721).
Engagement intention Variable Category Frequency Per cent

I intend to engage with fashion SMM this year. Age 18–29 years 159 22.1
I intend to engage with fashion SMM more frequently this year. 30–39 years 159 22.1
I intend to engage with fashion SMM for information or communication with people 40–49 years 159 22.1
this year. 50–59 years 158 21.9
I will not hesitate to engage with fashion SMM for information or communication 60–69 years 86 11.9
with people.
Gender Male 355 49.2
Media richness Female 360 49.9
When the social network site enables instant feedback, I engage more. Chose not to say 6 0.8
When the social network site provides rich and varied communication and response
Engagement hours per week 1–3 h 38 5.3
tools such as 'like', 'comment', 'post' or 'share', I engage more.
4–6 h 141 19.6
When the social network site enables a variety of message cues such as video, audio,
7–9 h 135 18.7
picture, or text, I engage more.
10–12 h 118 16.4
Content trustworthiness 13 h and more 289 40.0
Undependable vs. Dependable.
Education Undergraduate or lower 341 47.3
Dishonest vs. Honest.
Graduate 285 39.5
Unreliable vs. Reliable.
Postgraduate 9 13.2
Insincere vs. Sincere.
Untrustworthy vs. Trustworthy. Total – 721 100.0

Consumption Behaviour (items 1–4)

I read a poster related to fashion goods.
I watched a video/ picture/graphics related to fashion goods. 3.2. Sample and data
I followed a poster/picture/graphics related to fashion goods.
I clicked on a product link posted on my MESM to get more information. The primary data was collected via a survey questionnaire, ad-
Contribution Behaviour (items 5–8) ministered online by Qualtrics, a leading global data service company.
I commented on a poster/video/picture related to fashion goods. Our sample was sourced via Qualtrics's rigorous online sampling pro-
I shared/reposted/retweet a poster/video/picture/link initially posted on my MESM
cess from their actively managed market research panels. Survey re-
by the fashion brand.
I recommended fashion goods. spondents who are social media users were targeted. To ensure a
I “Liked”/“Disliked” a poster/video/picture/graphics related to fashion goods. sample with experience of engagement with social media with fashion
Creation Behaviour (items 9–12) SMM content, we designed one question to filter out those who did not
I initiated a discussion related to fashion goods. have such experience: 'Do you interact with fashion via social media, in
I posted a picture/graphics/video related to fashion goods. any of the following ways? Please tick all that apply'. The multiple-
I posted a selfie relating to a new fashion item. choice questions included seven items: 'Browsed fashion items',
I wrote a review related to fashion goods.
'Recommended fashion items’ 'Read or viewed fashion posts or images',
'Watched fashion videos', 'Followed Fashion brands', 'Posted pictures of
fashion items', and 'None of the above'. If 'None of the above' was
chosen, this terminated the survey, leaving an uncompleted ques-
tionnaire which was excluded from the data analysis. As a result, a total
advice from the extant literature on the advantages of including more sample of 721 valid responses was collated for analysis after data
items in the reflective measurement in Structural Equation Modeling cleaning.
(Hair, Sarstedt, Hopkins, & Kuppelwieser, 2014), we added the fourth Table 2 shows the demographic profile of the sample. Respondents
item, ‘I will not hesitate to engage …’, selected from Ayeh (2015). Five aged between 18 and 69 years were split equally on the base of gender
pairs of semantic phrases capturing content trustworthiness were also (male/female). All respondents in the final sample reported that they
adopted from Ayeh (2015). Based on Daft and Lengel (1986) and interacted with fashion SMM content on social media and that they
Brunelle (2009), the media richness was operationalised as three items: spent at least one hour per week on social networking sites. Forty per
feedback, communication tools, and message cues. Initially introduced cent of respondents reported that they spent over 13 h engaging with
by Daft and Lengel (1986), media richness was interpreted by in- social media each week, with 75 per cent spending more than seven
dicators for the selection of a medium (e.g. either an email or a phone hours per week.
call for communication) to effectively convey a message. However, this PLS-SEM (Partial Least Squares and Structural Equation Modelling)
study focuses on digital marketing communication and consumers’ was utilised for the data analysis. The most commonly used covariance-
engagement activities by using media communication tools such as based SEM (CB-SEM) was also considered but not adopted in this in-
'follow' and 'share'. Therefore, the three-item measurement used in this stance, for reasons explained below. Both CB-SEM and PLS-SEM allow
study were drawn on Brunelle (2009) study of perceived media rich- the analysis of structural relationships between latent constructs based
ness. on the SEM technique. However, the two SEM approaches vary in
We devised twelve items to measure engagement behaviours (see computation procedures, assumptions, and the assessment of structural
Table 1). As explained in section 2, consumption, contribution, and model fit (Hair et al., 2014). The rationale for the use of PLS-SEM is due
creation are the three levels of engagement behaviour, and the cate- to the advantages that PLS-SEM has over CB-SEM. PLS-SEM can be
gorisation of the three levels is based on the taxonomy of engagement easily applied to provide a solution to complex models, for instance,
activities by Schivinski et al. (2016). Items 1–4 describe the consumption models containing single-item constructs, a large number of constructs
activities, i.e., “read”, “watch”, “follow” and “click”, of consumer en- and hypotheses, and complex relationships such as mediation and or
gagement associated with fashion SMM content; items 5–8 describe the moderation (Hair et al., 2014; Astrachan, Petel, & Wanzenried, 2014).
contribution activities, i.e., “comment”, “share”, “recommend” and All of these conditions are met in this study.
“like/dislike”; items 9–12 describe the creation activities, i.e., “initiate a Moreover, the PLS-SEM approach offers the researcher the flex-
discussion”, “post a selfie”, “post a picture” and “write a review”, which ibility to introduce new latent variables and new relationships to an
are associated with fashion SMM content. established theory (Richter, Cepeda, Roldán, & Ringle, 2016), which

5
D. Cao, et al. Journal of Business Research xxx (xxxx) xxx–xxx

Table 3 discriminant validity between consumption and contribution, indicates a

Descriptive of the most engaged social media platforms in fashion. potential discriminant problem according to the HTMT .90 criterion.
Social media platforms Frequency Per cent Cumulative Percent However, the confidence interval, CI [0.884; 0.962], as shown in PLS3,
is between [2.5%, 97.5%], suggesting that these two constructs have no
Facebook 434 56.4 56.4 significant issue with discriminant validity. Therefore, the discriminant
Instagram 94 12.2 68.6
validity of all constructs was established (Henseler, Ringle, & Sarstedt,
YouTube 86 11.2 79.8
Pinterest 60 7.8 87.6
2015).
Twitter 57 7.4 95.0
Snapchat 6 0.8 95.8 4.3. Common method bias
Tumblr 6 0.8 96.6
Reddit 4 0.5 97.1
Our data was collected using a questionnaire, i.e. self-reported data
LinkedIn 2 0.3 97.4
Google+ 2 0.3 97.7 from a single source. We are aware that this may cause common
WhatsApp 1 0.1 97.8 method bias due to consistency motif and social desirability (Podsakoff,
Other 17 2.2 100.0 MacKenzie, Lee, & Podsakoff, 2003). To minimise the possibility, we
Total 769 100.0 100.0 followed the procedural remedies recommended by Podsakoff,
MacKenzie, and Podsakoff (2012). First, we eliminated the ambiguity of
item wording to get accurate answers from respondents to the ques-
also applies to this research. For instance, we examined 15 hypotheses tionnaire items (Podsakoff et al., 2003). Second, we conducted a post-
and the moderation effect of two contextual factors on the relationships hoc examination by employing the most commonly used method,
between the engagement intention and the three behavioural responses. Harmen's one-factor test. The result showed that four factors were
For these reasons, PLS-SEM was judged to be an appropriate choice for present, and the principal component extracted explained 39.78 per-
this study. cent of the variance of all 24 variables, which suggests that common
method bias is unlikely to be a serious concern for this study (Podsakoff
4. Results et al., 2003; Liang, Saraf, Hu, & Xue, 2007).

4.1. Descriptive statistics of social media platforms 4.4. Model and hypothesis testing

Table 3 reports consumers' most popular social media sites in a Next, we examined the structural relationships and the model fit.
fashion context. Facebook sits at the top of the list, used by 56.4 per Table 6 reports the results of the algorithm and bootstrapping tests
cent of the 759 participants. This result is no surprise due to the po- (based on 5,000 samples), including the effect sizes (coefficient β), the
pularity of Facebook, and its dominant position among social media corresponding significance of the effects (T values and ρ values) and
sites, However, instead of YouTube, Twitter, or LinkedIn, which are conclusions of support for each hypothesis (Yes or No).
widely reported to be the most popular social media sites, our data As shown in Table 6, engagement intention is significantly related to
indicates that following Facebook, Instagram and Pinterest are the all three levels of engagement behaviours with medium-size effects,
second and fourth most engaged social media platforms (with YouTube e.g., engagement intention - > consumption (β = 0.414, ρ < 0.001),
in third place). This is a slightly surprising result; however, it is likely to engagement intention - > contribution (β = 0.472, ρ < 0.001), and
reflect our particular interest in media richness and its role in influ- engagement intention - > creation (β = 0.462, ρ < 0.001). These
encing consumers' engagement online. This will be discussed further in findings support H1a, H1b, and H1c.
section 5 (Discussion). Our results also support the moderation effect of media richness.
First, the causal relationships between engagement intention and all
4.2. Construct reliability and validity three levels of engagement behaviours are statistically significant and
positive, with small-to-medium-size effects, e.g., media richness
Before conducting factor analysis, we tested the normality of each - > consumption (β = 0.280, ρ < 0.001), media richness - > con-
scale in our study. Appendix A reports the mean, standard deviation, tribution (β = 0.235, ρ < 0.001), and media richness - > creation
skewness, and kurtosis of all measures. Values of skewness and kurtosis (β = 0.123, ρ = 0.004). We find support for H2a, H2b, and H2c.
are all acceptable (Tabachnick & Fidell, 2007). Collinearity refers to the Second, the relationships between engagement intention (EI) and all
instance that the variables in the set are very highly correlated three levels of engagement behaviours are positively moderated by
(Tabachnick & Fidell, 2007, p. 88–90), which may cause statistical media richness (MR), with small-size effects: MR * EI - > consumption
problems. Therefore, we examined collinearity statistics (VIF) for our (β = 0.056, ρ = 0.038), MR * EI - > contribution (β = 0.113,
dataset, and results showed no evidence of such violation in our mea- ρ < 0.001), and MR * EI - > creation (β = 0.123, ρ < 0.001). These
sures, e.g., Inner VIF values ranging from 1.288 to 1.572, Outer VIF findings support H3a, H3b, and H3c. Hence, our results confirm that
values ranging from 1.436 to 2.931. The test results of normality and media richness has a positive effect on consumer engagement beha-
collinearity gave confidence in proceeding to the next step of factor viour in all three levels (consumption, contribution, and creation). In
analysis. addition, media richness significantly moderates the effect of con-
We first ran the PLS algorithm for factor analysis using SmartPLS 3. sumers' social media engagement intention on consumers’ actual be-
Table 4 displays the results for construct reliability and validity testing. haviour at all three levels.
All outer loadings, ranging from 0.748 to 0.890, are above the threshold Content trustworthiness is significantly related to consumption (H4a:
value of 0.70. All scores of Cronbach's α and Dijkstra-Henseler's rho β = 0.128, ρ < 0.001) and contribution (H4b: β = 0.095, ρ = 0.007)
(rho_A), ranging from 0.768 to 0.911, are above the threshold value of but insignificantly related to creation (H4c: β = 0.048, ρ = 0.199).
0.70, indicating consistency and reliability. All scores of Average Var- Hence, H4a and H4b are supported while H4c is not supported. Further,
iance Extracted (AVE) range from 0.612 to 0.745, suggesting the in- the moderation effects of content trustworthiness (CT) are all statisti-
ternal convergent validity of the constructs. cally insignificant, e.g., CT * EI - > consumption (β = -0.003,
We also checked the discriminant validity of the constructs using ρ = 0.902), CT * EI - > contribution (β = 0.002, ρ = 0.930), and CT *
Heterotrait-Monotrait Ratio Criterion (HTMT). All HTMT values EI - > creation (β = 0.038, ρ = 0.167). Therefore, H5a, H5b and H5c
(Table 5) except one are lower than the suggested value of 0.9 (Ringle, are not supported. Hence, our findings confirm that consumers’ trust in
Wende, & Becker, 2015). The exceptional one, related to the SMM content positively affects their consumption and contribution

6
D. Cao, et al. Journal of Business Research xxx (xxxx) xxx–xxx

Table 4
Construct reliability and validity.
Indicator Consumption Contribution Creation Engagement Intention Media Richness Content Trustworthiness

EAct1 0.761
EAct2 0.779
EAct3 0.832
EAct4 0.754
EAct5 0.856
EAct6 0.848
EAct7 0.825
EAct8 0.748
EAct9 0.879
EAct10 0.890
EAct11 0.864
EAct12 0.825
EI1 0.815
EI2 0.843
EI3 0.805
EI4 0.822
MR1 0.838
MR2 0.832
MR3 0.806
CT1 0.838
CT2 0.836
CT3 0.863
CT4 0.826
CT5 0.883
AVE 0.612 0.673 0.748 0.675 0.681 0.722
rho_A* 0.793 0.837 0.888 0.840 0.768 0.911
Crobach’s α 0.788 0.837 0.888 0.840 0.768 0.904

Note: *Dijkstra-Henseler’s rho (rho_A).

activities in social media communities. However, there is no evidence the context of fashion SMM. Results suggest that engagement intention
to support the effect of content trustworthiness on consumers’ creation is the key determinant of all three levels of consumer engagement be-
activities; the moderating role of content trustworthiness is not statis- haviours, with medium-size effects. This result is consistent with most
tically significant. TPB empirical studies (e.g., Cooke & French, 2008). For instance, Taylor
Table 6 also reports the model fit indicators, e.g., SRMR (0.110), and Todd (1995) studied the intention to act on information technology
Chi-square (2,033.76), and NFI (0.799). Fig. 2 displays R square values usage and found a comparable medium-size effect (0.38). Similarly, Al-
for three dependent variables, i.e. consumption (0.429), contribution Debei et al. (2013) examined the relationship between continual par-
(0.415), and creation (0.287). When we make a comparison of these ticipation intentions and behaviour on Facebook and identified a
results with similar studies that have empirically examined the asso- medium-size effect (0.414).
ciation between intention and actual behaviour but without moderation Our results also suggest that the richness of social media is positively
consideration in their models, we find that the R square values of past related to consumer engagement behaviours at all three levels – con-
studies - e.g. 0.296 (Cao et al., 2019), and ranging from 0.35 to 0.39 sumption, contribution, and creation. The result supports a number of
(Venkatesh et al., 2003) - are lower than the R square values reported similar studies (e.g., Kim et al., 2017; Guidry et al., 2018). Guidry et al.
here, which range from 0.287 to 0.429. This suggests that SMEB, in (2018) investigated in their study how posts of refugees on two visual-
terms of consumption, contribution, and creation, is well explained by the based social media platforms, namely Instagram and Pinterest, differ in
model, and by the moderators in particular. terms of content and engagement; their results suggest that the majority
of the humanitarian- and security-concern variables presented a sig-
5. Discussion nificant association with public engagement on one social media plat-
form over another. For instance, security-concern expressions sig-
This research investigates the moderation effects of two social nificantly associate with Pinterest, while humanitarian-concern
media contextual factors, namely, media richness and content trust- expressions were found to associate with Instagram. Further, the study
worthiness, between engagement intention and engagement behaviour in revealed that the association of public engagement behaviour with

Table 5
Heterotrait-Monotrait Ratio (HTMT) and Confidence Interval.
1 2 3 4 5

1.Consumption
2.Content trustworthiness 0.460
CI[0.379;0.537]
3.Contribution 0.925 0.419
CI [0.884;0.962] CI[0.344;0.497]
4.Creation 0.732 0.309 0.889
CI[0.679;0.781] CI[0.231;0.384] CI[0.849;0.920]
5.Engagement intention 0.727 0.531 0.702 0.565
CI[0.656;0.791] CI[0.445;0.615] CI[0.634;0.758] CI[0.481;0.629]
6.Media richness 0.645 0.343 0.565 0.378 0.628
CI[0.559;0.714] CI[0.248;0.434] CI[0.477;0.635] CI[0.289;0.455] CI[0.547;0.701]

7
D. Cao, et al. Journal of Business Research xxx (xxxx) xxx–xxx

Table 6
Results of algorithm and bootstrapping tests.
Hypothesis Coefficient T-value ρ - value Support

H1a: Engagement intention → Consumption 0.414 10.835 0.000 Yes

H1b: Engagement intention → Contribution 0.472 12.761 0.000 Yes
H1c: Engagement intention → Creation 0.463 11.454 0.000 Yes
H2a: Media richness → Consumption 0.280 7.152 0.000 Yes
H2b: Media richness → Contribution 0.235 6.029 0.000 Yes
H2c: Media richness → Creation 0.123 2.874 0.004 Yes
H3a: MR * EI → Consumption 0.056 2.070 0.038 Yes
H3b: MR * EI → Contribution 0.113 4.656 0.000 Yes
H3c: MR * EI → Creation 0.123 5.526 0.000 Yes
H4a: Content trustworthiness → Consumption 0.128 3.656 0.000 Yes
H4b: Content trustworthiness → Contribution 0.095 2.693 0.007 Yes
H4c: Content trustworthiness → Creation 0.048 1.285 0.199 No
H5a: CT * EI → Consumption -0.003 0.123 0.902 No
H5b: CT * EI → Contribution 0.002 0.088 0.930 No
H5c: CT* EI → Creation 0.038 1.381 0.167 No

Model fit summary: SRMR = 0.110; Chi-square = 2,033.763; NFI = 0.799; Consumption (R2 = 0.429); Contribution (R2 = 0.415); Creation (R2 = 0.287).

Note: MR = Media richness; EI = Engagement intention; CT = Content trustworthiness. Significant at 0.05 (two-tailed).

different social media platforms is related to the variation of the media Moreover, our results suggest that media richness significantly
richness of the two platforms. For instance, Instagram displayed more moderates the intention-behaviour relationship. The moderation effect
episodic framing, while Pinterest contained more thematic framing. of media richness has rarely been explored in the existing literature.
Extant literature offers some insights as to why media richness might Therefore, there is a need for further study, particularly in different
affect consumer engagement, i.e. the opportunity to engage in online industries or sectors other than fashion which is the research context for
brand communities or social-media based communities (Schau et al., this study.
2009; Wirtz et al., 2013; Dessart et al., 2015). Media richness facilitates In contrast to previous studies (e.g., Kumar et al., 2016; Wang et al.,
and enables social media users to enjoy social interactions with others 2017) using a single-medium context or examining a single behavioural
sharing common interests within the social media environment (Kuo & response, we examined the richness of all possible social media used by
Feng, 2013). However, our results also present some differences from respondents. This approach enhances our methodological robustness
the extant literature. For instance, according to Pentina et al. (2018), and builds a more substantial theoretical contribution. By using the
luxury consumers' SMEB is not related to the choice of social media multiple social media setting, we are able to conceptualise and oper-
platforms. ationalise media richness in terms of a variety of message cues, a

Fig. 2. SmartPLS output of the algorithm test.

8
D. Cao, et al. Journal of Business Research xxx (xxxx) xxx–xxx

variety of communication tools and instantaneous feedback. Also, we contribution, and creation associated with SMM content. We also high-
examined multiple behavioural responses rather than the single beha- lighted the moderation effect of social media contextual factors,
vioural response used in most extant literature. This contributes to the namely, media richness and content trustworthiness, on the relation-
advancement of TPB. The multiple behaviours setting in our study ex- ship between engagement intention and engagement behaviour. By
plain not only users' decision to engage but also the depth of engage- doing so, we address a gap in the extant literature and contribute to the
ment the users develop. The latter is critical for social media marketers continuous development of behavioural theory, particularly TPB, in the
to build up their retailing and brand performance. Further, our research contemporary landscape of social media. Second, the outcome of this
supports the moderation effect of media richness; the relationship be- research offers an enhanced understanding of consumer engagement
tween engagement intention and action becomes stronger when a social with fashion SMM, a topical issue given the increasingly interactive and
medium is “richer” rather than “leaner”. This is a novel finding and an experiential nature of consumer-brand relationships in the digital
essential contribution to the knowledge and understanding of SMEB marketing era (Hollebeek et al., 2014; Dessart et al., 2015). This study
and the various SMEB levels. offers insights into how social media contextual factors such as media
We also found that content trustworthiness affects consumption and richness and content trustworthiness can support the extension of
contribution. It is understandable that trustworthy content encourages consumers' scope of engagement beyond core purchase or transactional
social media users to engage in activities that include making connec- relationships. As studies have demonstrated (Hollebeek et al., 2014;
tions with other social media users such as “follow”, “share”, “com- Vargo & Lusch, 2004), by being engaged, consumers are likely to de-
ment” and “recommend”. This supports the results of earlier studies velop a more profound notion of brand association. Third, this research
(Swani et al., 2017; Zhang & Gupta, 2018). Swani et al. (2017) provide focuses on the role of online social media context played in SMEB.
insights into the management of the content trustworthiness of SMM, Based on media richness theory and source credibility theory, we se-
suggesting that social media marketers use product brand names, lected two factors, media richness and content trustworthiness, as the
emotional appeals, and direct-calls-to-purchase in their B2C Facebook proxy of the study. Online social media, as a research context, has been
posts to engage end-users with Facebook “likes” and “comments”; increasingly popular in recent decades. However, understanding the
corporate brand names, functional appeals, and links or cues for in- setting itself and its role is yet minimal. This research, as one of the first
formation search are used in their B2B Facebook posts for more effec- studies, contributes to the understanding of the social media context.
tive engagement of their business users. Taking these insights further,
we distinguish between the different levels of consumer behaviour 6.2. Practical implications
within the SMM community, suggesting that content trustworthiness
does not influence the highest level of consumer engagement beha- Our empirical results provide useful insights for fashion brands
viour, i.e. creation. Although an in-depth explanation of this result is seeking to enhance consumer engagement with their SMM through
beyond the scope of this research, we postulate that the creation of managing content trustworthiness and media richness. Engagement
(new) content is independent of the trustworthiness of the original intention is a critical factor in determining consumers' engagement
SMM message. activities. However, to what extent the intention affects actual con-
In contrast to media richness, we found that there is no evidence to sumer engagement is significantly moderated by social media con-
support the moderation effect of content trustworthiness on any en- textual factors, i.e. media richness and content trustworthiness.
gagement behaviour. One plausible explanation could be that when Therefore, to achieve the most effective outcome from the use of SMM
social observation, social networking, or social identity become the as a marketing strategy, the careful selection of media (for its richness)
main motives for users to engage in social media (Oliveira et al., 2016), and careful generation of digital marketing contents should not be ne-
what appeals to consumers most in enacting their engagement intention glected.
is the appeal of social networking itself rather than its embedded con- Social media sites with instant feedback can encourage interaction.
tent. Under such circumstances, users have a greater interest in inter- In particular, social media sites such as Facebook, Twitter and
acting and socialising with other community members than they do in YouTube, with richer response tools, can enhance consumer engage-
content trustworthiness. ment activities with fashion SMM content, using 'following', 'sharing',
'commenting', 'retweeting', 'posting' and 'reviewing'. Social media sites
6. Conclusion such as Instagram and Pinterest, with richer message cues in the form of
video, audio, and high-quality pictures, can engage more fashion con-
To sum up, media richness, as a social media contextual factor, sumers, and this is reflected in our research. However, it is important to
positively affects various levels of consumers' engagement – consump- be aware that each social media site has its own unique ‘flavour’ of
tion, contribution, and creation behaviours. Moreover, media richness richness; fashion social media marketers have to select a particular
moderates the relationship between engagement intention and beha- social media site that serves their specific marketing strategy most ef-
viour. The consumer is more likely to engage with SMM when media fectively. For instance, to encourage communication and interaction
platforms are richer rather than leaner. Similarly, content trustworthi- around the design ideas behind fashion products with social media
ness, as another social media contextual cue, also positively affects consumers, fashion marketers could consider sites such as Facebook and
consumers' consumption and contribution behaviour. However, content Twitter with instant feedback and live interaction. However, if they aim
trustworthiness does not appear to have a significant influence on to launch a new product and encourage consumption, social media sites
creation behaviour. The moderation effect of content trustworthiness on such as Instagram and Pinterest with higher-quality pictures and video
the relationship between engagement intention and behaviour was also could be prioritised. If the goal is to reach and inform a wider audience
found to be insignificant. of consumers, the more ‘traditional’ and popular social media sites such
as Facebook may remain a wise choice.
6.1. Theoretical contributions Social media marketers should also seek to enhance consumer en-
gagement with SMM by building up the trustworthiness of SMM content
The interactive nature of the social media context brings new di- embedded within their social network communities. To build up the
mensions to the relationship between engagement intention and en- trustworthiness of fashion marketer-generated content, instead of using
gagement behaviour. Our first contribution is that we extend knowl- emotion-focused appeals (e.g. using peers' word of mouth, or celebrity
edge of TPB by exploring multiple behavioural responses rather than a endorsements), fashion social media marketers should respect con-
single action. We examined the effect of social media engagement in- sumers' cognitive judgement and rational choices by focusing more on
tention on three levels of actual engagement behaviours: consumption, facts and features of their products or services when generating their

9
D. Cao, et al. Journal of Business Research xxx (xxxx) xxx–xxx

SMM content. That is, informative, dependable, reliable and trust- dimensions and their conceptualisation could be a future research di-
worthy social media content can work to build up trustworthiness and rection. Second, alongside consumer engagement, engagement with
hence consumers' social media engagement. Moreover, although peers other stakeholders such as investors, local or national governments, and
may be viewed as personally trustworthy, peer-generated content can the broader public on online social media is also crucial for the per-
be seen as less trustworthy due to their lack of expertise. However, formance management of firms (Yang, Cao, Andrikopoulos, Yang, &
messages from social media opinion leaders are more trustworthy due Bass, 2020). This to date has been the subject of little research, and it
to their knowledge of specific products or services. Therefore, to choose therefore provides another research direction in the future. Third, it
the latter as the source of SMM may engage consumers more effectively. should be noted that our findings are based in the context of fashion
social media, and care should be taken for the generalisation to other
6.3. Limitations and future research industries or sectors. This also indicates another future research direc-
tion.
There are several limitations to our study, which are viewed as of-
fering the potential for future research. First, we used two fundamental Acknowledgement
constructs, trustworthiness and media richness, as proxies for a range of
contextual factors. However, the media context is complex and multi- This work was supported by Coventry University. The authors ac-
dimensional and, therefore, systematic exploration of the context knowledge the financial support of the University.

Appendix A. Assessment of normality

Mean Std. Deviation Skewness Kurtosis

Statistic Statistic Statistic Std. Error Statistic Std. Error

Media Richness1 3.95 0.863 −0.627 0.091 0.415 0.182

Media Richness2 4.30 0.862 −1.346 0.091 1.895 0.182
Media Richness3 4.17 0.868 −1.024 0.091 0.951 0.182
Engagement Act1 2.99 1.247 −0.136 0.091 −0.929 0.182
Engagement Act2 3.48 1.248 −0.462 0.091 −0.766 0.182
Engagement Act3 3.30 1.268 −0.340 0.091 −0.889 0.182
Engagement Act4 3.49 1.233 −0.518 0.091 −0.686 0.182
Engagement Act5 2.75 1.393 0.216 0.091 −1.219 0.182
Engagement Act6 2.63 1.417 0.283 0.091 −1.274 0.182
Engagement Act7 2.64 1.364 0.248 0.091 −1.190 0.182
Engagement Act8 3.43 1.341 −0.491 0.091 −0.940 0.182
Engagement Act9 2.22 1.288 0.692 0.091 −0.715 0.182
Engagement Act10 2.33 1.365 0.611 0.091 −0.920 0.182
Engagement Act11 2.08 1.353 0.935 0.091 −0.463 0.182
Engagement Act12 2.14 1.336 0.850 0.091 −0.589 0.182
Content trustworthiness1 3.85 0.934 −0.780 0.091 0.682 0.182
Content trustworthiness2 3.83 0.909 −0.604 0.091 0.315 0.182
Content trustworthiness3 3.85 0.908 −0.552 0.091 0.146 0.182
Content trustworthiness4 3.73 0.990 −0.522 0.091 −0.001 0.182
Content trustworthiness5 3.79 0.931 −0.552 0.091 0.277 0.182
Engagement intention1 4.06 0.906 −1.025 0.091 1.126 0.182
Engagement intention2 3.47 1.020 −0.202 0.091 −0.459 0.182
Engagement intention3 3.98 0.881 −0.969 0.091 1.279 0.182
Engagement intention4 3.64 1.062 −0.543 0.091 −0.209 0.182

References brand community: An exploratory research. Journal of Business Research, 66(1),

105–114.
Bronner, F., & Neijens, P. (2006). Audience experiences of media context and embedded
Ajzen, I. (1991). The theory of planned behavior. Organisational Behavior and Human advertising: A comparison of eight media. International Journal of Market Research.
Decision Processes, 50(2), 179–211. 48(1), 81–100.
Ajzen, I., & Fishbein, M. (1980). Understanding Attitudes and Predicting Social Behavior. Brunelle, E. (2009). Introducing media richness into an integrated model of consumers'
Englewood Cliffs, NJ: Prentice-Hal. intentions to use online stores in their purchase process. Journal of Internet Commerce,
Ajzen, I., & Sheikh, S. (2013). Action versus inaction: Anticipated effect in the theory of 8(3), 222–245.
planned behavior. Journal of Applied Social Psychology, 43(1), 155–162. Cao, D., Tao, H., Wang, Y., Tarhini, A., & Xia, S. (2019). Acceptance of automation
Al-Debei, M., Al-Lozi, E., & Papazafeiropoulou, A. (2013). Why people keep coming back manufacturing technology in china: An examination of perceived norm and organi-
to Facebook: Explaining and predicting continuance participation from an extended sational efficacy. Production Planning and Control, 660–672.
theory of planned behaviour perspective. Decision Support Systems, 55(1), 43–54. Carlson, J. R., & Zmud, R. W. (1999). Channel expansion theory and the experiential
Astrachan, C. B., Patel, V. K., & Wanzenried, G. (2014). A comparative study of CB-SEM nature of media richness perceptions“. Academy of Management Journal, 42(2),
and PLS-SEM for theory development in family firm research. Journal of Family 153–170.
Business Strategy, 5(1), 116–128. Chang, Y., Yu, H., & Lu, H. (2015). Persuasive messages, popularity cohesion, and mes-
Ayeh, J. K. (2015). Travellers' acceptance of consumer-generated media: An integrated sage diffusion in social media marketing. Journal of Business Research, 68(4),
model of technology acceptance and source credibility theories. Computers in Human 777–782.
Behavior, 48, 173–180. Cooke, R., & French, D. (2008). How well do the theory of reasoned action and theory of
Bail, C. (2016). Cultural carrying capacity: Organ donation advocacy, discursive framing, planned behaviour predict intentions and attendance at screening programmes? A
and social media engagement. Social Science & Medicine, 165, 280–288. meta-analysis. Psychology & Health, 23(7), 745–765.
BBC (2018). Only one if four trust social media, says survey. Accessed 7th August 2018 Daft, R. L., & Lengel, R. H. (1986). Organisational information requirements, media
https://www.bbc.co.uk/news/business-42771740. richness and structural design. Management Science, 32(5), 554–571.
Bendoni, W. (2017). Social media for fashion marketing: Storytelling in a digital world. Dahlen, M. (2005). The medium as a contextual cue: Effects of creative media choice.
London: Bloomsbury Visual Arts. Journal of Advertising. 34(3), 89–98.
Braten, I., Braasch, J., Strømsø, H., & Ferguson, L. (2015). Establishing trustworthiness Dessart, L. (2017). Social media engagement: A model of antecedents and relational
when students read multiple documents containing conflicting scientific evidence. outcomes. Journal of Marketing Management, 33(5–6), 375–399.
Reading Psychology, 36(4), 315–349. Dessart, L., Veloutsou, C., & Morgan-Thomas, A. (2015). Consumer engagement in online
Brodie, J. R., Ilic, A., Juric, B., & Holleneek, L. (2011). Consumer engagement in a virtual brand communities: A social media perspective. Journal of Product & Brand

10
D. Cao, et al. Journal of Business Research xxx (xxxx) xxx–xxx

Management, 24(1), 28–42. remedies. Journal of Applied Psychology, 88, 879–903.

Fulgoni, G. M. (2015). How brands using social media ignite marketing and drive growth. Podsakoff, P. M., MacKenzie, S. B., & Podsakoff, N. P. (2012). Sources of method bias in
Journal of Advertising Research, 55(3), 232–236. social science research and recommendations on how to control it. Annual Review of
Fulgoni, G., & Lipsman, A. (2017). The downside of digital word of mouth and the pursuit Psychology, 63, 539–569.
of media quality: How social sharing is disrupting digital advertising models and Popovic, K. (2016). Satellite Marketing: Using Social Media to Create Engagement. London;
metrics. Journal of Advertising Research, 57(2), 127–131. New York: CRC Press.
Gefen, D. (2002). Reflections on the dimensions of trust and trustworthiness among on- Prendergast, G., Liu, P., & Poon, D. (2009). A Hong Kong study of advertising credibility.
line consumers. Database for Advances in Information Systems, 33(3), 38–53. Journal of Consumer Marketing, 26(5), 320–329.
Guidry, J., Austin, L. L., Carlyle, K. E., Freberg, K., Cacciatore, M., Jin, Y., & Messner, M. Rapp, A., Beitelspacher, L., Grewal, D., & Hughes, D. (2013). Understanding social media
(2018). Welcome or not: Comparing refugee posts on Instagram and Pinterest. effects across seller, retailer, and consumer interactions. Journal of Academy of
American Behavioral Scientist, 62(4), 512–531. Marketing Science, 41(5), 547–566.
Guidry, J., Jin, Y., Orr, C. A., Messner, M., & Meganck, S. (2017). Ebola on Instagram and Richard, P. S. (2015). Domain Analysis for Knowledge Organization. Elsevier. https://doi.
Twitter: How health organisations address the health crisis in their social media org/10.1016/C2014-0-02303-9.
engagement. Public Relations Review, 43(3), 477–486. Richter, N. F., Cepeda, G., Roldán, J. L., & Ringle, C. M. (2016). European management
Habibi, M. R., Laroche, M., & Richard, M. (2016). Testing an extended model of consumer research using partial least squares structural equation modeling (PLS-SEM).
behavior in the context of social media-based brand communities. Computers in European Management Journal, 34(6), 589–597.
Human Behavior, 62, 292. Ringle, C. M., Wende, S., & Becker, J.-M. (2015). SmartPLS 3. Boenningstedt: SmartPLS
Hair, J., Jr, Sarstedt, M., Hopkins, L., & Kuppelwieser, V. G. (2014). Partial least squares GmbH http://www.smartpls.com.
structural equation modeling (PLS-SEM): An emerging tool in business research. Sandberg, T., & Conner, M. (2008). Anticipated regret as an additional predictor in the
European Business Review, 26(2), 106–121. theory of planned behavior: A meta-analysis. British Journal of Social Psychology, 47,
Henseler, J., Ringle, C., & Sarstedt, M. (2015). A new criterion for assessing discriminant 589–606.
validity in variance-based structural equation modeling. Journal of the Academy of Schau, H. J., Muñiz, A. M., Jr., & Arnould, E. J. (2009). How brand communities create
Marketing Science, 43(1), 115–135. value. Journal of Marketing, 73(5), 30–51.
Hollebeek, L., Glynn, M. S., & Brodie, R. J. (2014). Consumer brand engagement in social Schivinski, B., Christodoulides, G., & Dabrowski, D. (2016). Measuring consumers' en-
media: Conceptualisation, scale development and validation. Journal of Interactive gagement with brand-related social- media content: Development and validation of a
Marketing, 28(2), 149–165. scale that identifies levels of social- media engagement with brands. Journal of
Hovland, C. I., & Weiss, W. (1951). The influence of source credibility on communication Advertising Research, 56(1), 64–80.
effectiveness. Public Opinion Quarterly, 15, 635–650. Swani, K., Milne, G. R., Brown, B. P., Assaf, A. G., & Donthu, N. (2017). What messages to
Hwong, Y., Oliver, C., Van Kranendonk, M., Sammut, C., & Seroussi, Y. (2017). What post? Evaluating the popularity of social media communications in business versus
makes you tick? The psychology of social media engagement in space science com- consumer markets. Industrial Marketing Management, 62(April), 72–87.
munication. Computers in Human Behavior, 68, 480–492. Tabachnick, B. G., & Fidell, L. S. (2007). Using multivariate statistics (4th ed.). London:
Kahai, S. S., & Cooper, R. B. (2003). Exploring the core concepts of media richness theory: Pearson Education Inc.
The impact of cue multiplicity and feedback immediacy on decision quality. Journal Taylor, S., & Todd, P. A. (1995). Understanding information technology usage: A test of
of Management Information Systems, 20(1), 263–299. competing models. Information Systems Research, 6(2), 144–176.
Kaplan, A. M., & Haenlein, M. (2010). Users of the world, unite! The challenges and Teng, S., Khong, K., & Goh, W. (2015). Persuasive communication: A study of major
opportunities of social media. Business Horizons, 53(1), 59–68. attitude-behavior theories in a social media context. Journal of Internet Commerce,
Khan, M. (2017). Social media engagement: What motivates user participation and 14(1), 42–64.
consumption on YouTube? Computers in Human Behavior, 66, 236–247. Tormala, Z., & Clarkson, J. (2007). Assimilation and contrast in persuasion: The effects of
Khobzi, H., Lau, R., & Cheung, T. (2019). The outcome of online social interactions on source credibility in multiple message situations. Personality and Social Psychology
Facebook pages. Internet Research, 29(1), 2–23. Bulletin, 33(4), 559–571.
Kim, A., & Ko, E. (2012). Do social media marketing activities enhance customer equity? Vargo, S. L., & Lusch, R. F. (2004). Evolving to a new dominant logic for marketing.
An empirical study of luxury fashion brand. Journal of Business Research, 65(10), Journal of Marketing, 68(1), 1–17.
1480–1486. Venkatesh, V., & Bala, H. (2008). Technology acceptance model 3 and a research agenda
Kim, D. H., Seely, N. K., & Jung, J. H. (2017). Do you prefer, Pinterest or Instagram? The on interventions. Decision Sciences, 39(2), 273–315.
role of image-sharing SNSs and self-monitoring in enhancing ad effectiveness. Venkatesh, V., Morris, M. G., Davis, G. B., & Davis, F. D. (2003). User acceptance of
Computers in Human Behavior, 70, 535–543. information technology: Toward a unified view. MIS Quarterly: Management
Kim, Sung-Bum, & Kim, Dae-Young (2014). The effects of message framing and source Information Systems, 27(3), 425–478.
credibility on green messages in hotels. Cornell Hospitality Quarterly, 55(1), 64–75. Wang, R., Kim, J., Xiao, A., & Jung, Y. J. (2017). Networked narratives on humans of New
Kumar, A., Bezawada, R., Rishika, R., Janakiraman, R., & Kannan, P. K. (2016). From York: A content analysis of social media engagement on Facebook. Computers in
social to sale: The effects of firm-generated content in social media on customer Human Behavior, 66, 149–153.
behavior. Journal of Marketing, 80(1), 7–25. Wang, Y., & Yu, C. (2017). Social interaction-based consumer decision-making model in
Kuo, Y. F., & Feng, L. H. (2013). Relationships among community interaction char- social commerce: The role of word of mouth and observational learning. International
acteristics, perceived benefits, community commitment and oppositional brand loy- Journal of Information Management, 37(3), 179–189.
alty in online brand communities. International Journal of Information Management, Wang, C., Zhang, J., Yu, P., & Hu, H. (2018). The theory of planned behaviour as a model
33(6), 948–962. for understanding tourists' responsible environmental behaviours: The moderating
Lai, I., & Liu, Y. (2020). The effects of content likeability, content credibility, and social role of environmental interpretations. Journal of Cleaner Production, 194, 425–434.
media engagement on users' acceptance of product placement in mobile social net- Wirtz, J., den Ambtman, A., Bloemer, J., Horváth, C., Ramaseshan, B., Van De Klundert,
works. Journal of Theoretical and Applied Electronic Commerce Research, 15(3), 1–19. J., ... Kandampully, J. (2013). Managing brands and consumer engagement in online
Liang, H., Saraf, N., Hu, Q., & Xue, Y. (2007). Assimilation of enterprise systems: The brand communities. Journal of Service Management, 24(3), 223–244.
effect of institutional pressures and the mediating role of top management. MIS Yang, X., Cao, D., Andrikopoulos, P., Yang, Z., & Bass, T. (2020). Online social networks,
Quarterly, 31(1), 59–89. media supervision and investment efficiency: An empirical examination of Chinese
Lim, J. S., Hwang, Y., Kim, S., & Biocca, F. A. (2015). How social media engagement leads listed companies. Technological Forecasting & Social Change, 154, 119969.
to sports channel loyalty: Mediating roles of social presence and channel commit- Zhang, Z., & Gupta, B. B. (2018). Social media security and trustworthiness: Overview and
ment. Computers in Human Behavior, 46(1), 158–167. new direction. Future Generation Computer Systems, 86, 914–925.
Malthouse, E., & Hofacker, C. (2010). Looking back and looking forward with interactive Zhang, X. A., Borden, J., & Kim, S. (2018). Understanding publics' post-crisis social media
marketing. Journal of Interactive Marketing, 24(3), 181–184. engagement behaviors: An examination of antecedents and mediators. Telematics and
Mintel. (2018a). Beauty influencers and educators – UK – August 2018. http://academic. Informatics, 35(8), 2133–2146.
mintel.com/display/909862/. Accessed 17th September 2018.
Mintel. (2018b). Attitudes towards fashion – China – July 2018. http://academic.mintel. Dongmei Cao (PhD) PhD, is a Lecturer and Research Associate at the Faculty of Business
com/display/904466/. Accessed 17th September 2018. & Law, Coventry University, UK. Dongmei holds a PhD in Strategic Management from
Muntinga, D., Moorman, M., & Smit, E. (2011). Introducing COBRAs: Exploring moti- Coventry University. Her primary research interest is sustainability, which incorporates
vations for brand-related social media use. International Journal of Advertising, 30(1), sustainable performance management, technology applications, innovation, big data,
13–46. social media, and responsible consumption. Her research has been published in a few
Muralidharan, S., & Men, L. (2015). How peer communication and engagement motiva- domain journals such as Journal of Business Research, Technological Forecasting & Social
tions influence social media shopping behavior: Evidence from China and the United Change, and Production Planning & Control. She is an active reviewer for some interna-
States. Cyberpsychology, Behavior and Social Networking, 18(10), 595–601. tional journals and book publishers.
Oliveira, M. J., Huertas, M. K. Z., & Lin, Z. (2016). Factors driving young users' en-
gagement with Facebook: Evidence from Brazil. Computers in Human Behavior, 54,
54–61. Maureen Meadows is Professor of Strategic Management at the Centre for Business in
Society (CBiS) at Coventry University, UK. Maureen’s research interests include the use of
Pentina, I., Guilloux, V., & Micu, A. (2018). Exploring social media engagement behaviors
in the context of luxury brands. Journal of Advertising, 47(1), 55–69. strategy tools such as scenario planning and visioning by senior managers, and the post-
Phan, M., & Park, S. (2014). Introduction: Social media marketing and luxury brands. merger integration phase following M&A deals. With a background in mathematics,
Journal of Global Fashion Marketing: Social Media Marketing and Luxury Brands, 5(3), statistics and operational research, Maureen has over 20 years of experience of working
195–196. with 'big data' and customer analytics, both as a practitioner in the financial services
sector and an academic. She has published on the progress and problems experienced by
Podsakoff, P. M., MacKenzie, S. B., Lee, J.-Y., & Podsakoff, N. P. (2003). Common method
biases in behavioral research: A critical review of the literature and recommended organisations working on strategic projects such as market segmentation, relationship

11
D. Cao, et al. Journal of Business Research xxx (xxxx) xxx–xxx

marketing and customer relationship management. She is an active reviewer for some international journals and book publishers.

Donna Wong (PhD) is currently working as Associate Professor at the Graduate School of Senmao Xia (PhD) is a Senior Lecturer at School of Strategy and Leadership &
Sport Sciences, Waseda University, Japan. Her research interests relate to various man- International Center for Transformational Entrepreneurship, Coventry University, UK. His
agerial aspects of sports, such as digital media and sports broadcasts, and consumption main research interests include international knowledge transfer, returnee en-
behaviour in sports as well. Dr Wong has disseminated her works widely in books, trepreneurship and OFDI. He acts as guest editor for Technovation special issue &
chapters and international journals, in both the field of sports and general management. International Journal of Technology Management.

12