Since the social commerce concept is broad and is associated with several inconsistencies, including activities of e-commerce that involve social networks and social media, it currently lacks a standard definition (Sohn and Kim, 2020). According to Stephen and Toubia (2010), social commerce is a type of Internet-based social network that enables people to directly engage in the selling and marketing of products or services via websites and online communities. Social commerce is a term used to define an e-commerce practice involving social media and social networks (Liang et al., 2011; Mamonov and Benbunan-Fich, 2017). In addition, social commerce entails using Web 2.0 technologies to promote online social engagements and customer experience to help buyers purchase goods and services (Kim and Park, 2013; Yang et al., 2021). Sohn and Kim (2020) also define social commerce as “selling goods through SNSes as well as the electronic commerce based on a specific site.” We use the following definition of social commerce by Meilatinova (2021) in this study, “as an extension of e-commerce sites, integrated with social media and Web 2.0 technology to encourage online purchases and interactions with customers before, during, and after the purchase.”

The literature has identified two kinds of social commerce. The first is social networking websites incorporating tools to facilitate online transactions. The second is typical e-commerce websites incorporating elements such as like, share, and comment to encourage social networking (Liang et al., 2011; Zhang and Benyoucef, 2016; Hajli et al., 2017; Mou and Benyoucef, 2021). However, other social commerce categories can be established based on the combination of characteristics (Molinillo et al., 2021). Based on how consumer interactions are built, what their aims are, what information they exchange, and the sort of platform they utilise, Han et al. (2018), for example, identified eight distinct forms of social commerce.

A lot of attention has been given to social commerce (Lin et al., 2017; Mou and Benyoucef, 2021). Issues such as engagement in SC websites and sellers (Wongkitrungrueng and Assarut, 2020), information sharing (Hajli et al., 2017), value co-creation (Tajvidi et al., 2021), decision making in SC regarding purchase (Chen et al., 2017; Lãzãroiu et al., 2020), and the role of peers relations and social influence (Yang, 2021) are investigated in prior research. Electronic word of mouth (eWOM) (Cheng et al., 2021) and features of social commerce Web design (Molinillo et al., 2021) are also discussed in previous studies. However, trust and its formation have been paid little attention to in previous SC research (Leong et al., 2020, 2021; Meilatinova, 2021; Mou and Benyoucef, 2021).

According to social-technical theory (Figure 1), a system may be divided into social and technical subsystems (Bostrom and Heinen, 1977). In addition to this, the procedures and technologies make the technical subsystem, and these systems allow users to perform multiple functions. For example, users can convert inputs into outputs and do some predefined system activities. On the other hand, the social subsystem consists of knowledge, users, connections, values, and a mechanism of reward (Lin et al., 2019). The subsystems mentioned above are not isolated; instead, they are related and should function in tandem for optimal performance of the system (Bostrom and Heinen, 1977; Bostrom et al., 2009). According to this viewpoint, people and technology play essential parts in determining the functionality of the system since they mainly define the connection between the two subsystems (Kong et al., 2020).

This theory is relevant to social commerce, directed by web technology and the people involving the aforementioned subsystems (Kong et al., 2020). The social commerce subsystem consists of people (social media users, sellers, and buyers) and organisational components (groups, communities, and fan pages) supporting the social relationships and interactions, facilitating social support, and making people rely on and commit to an organisation (Bostrom et al., 2009; Liang et al., 2011; Wang and Zhang, 2012; Ng, 2013; Lin et al., 2019). The technological subsystem of social commerce comprises Web technology, including social media, social commerce elements, and e-commerce sites and process elements that support the SC operations like commercial information exchange and consumer evaluations (Liang et al., 2011; Wang and Zhang, 2012).

In addition, technological variations do not restrict the technical system and do not influence the system and the users of the social structure, as stated by the social-technical theory (Lin et al., 2019). E-commerce sites alter the technological environment by including SC features and social networking sites. As a result, the affordances that allow customers to do new activities are provided (i.e., new customer behaviours like information sharing about the product). Moreover, such modifications can also influence people and the structure of the task (i.e., examining how customers engage on online e-commerce sites and online platforms) (Bostrom et al., 2009). This subsystem also forms the basis for the social subsystem in social commerce. Finally, the internet and online sites also allow individuals to connect, build, and sustain social ties.

In contrast, progression in the technical subsystem of social commerce may be triggered by the social subsystem because organisations and people require valuable support, interactions, and social relationships. This can be obtained through technological capabilities (Kong et al., 2020). These two subsystems have been demonstrated to operate well together in social commerce (Lin et al., 2019). As a result, these two subsystems benefit customers by enhancing their shopping experience and purchasing decisions (Lin et al., 2019). Hence, the social-technical theory is an ideal theoretical framework since it may aid in understanding how many variables interact to build the trust of customers in social commerce. Using social–technical theory, it has been suggested that technology and people are the fundamental elements that propel social commerce and can truly depict the social commerce subsystems (Lin et al., 2019; Kong et al., 2020).

This study examines a social commerce platform setting that combines social networking sites with conventional e-commerce websites. Moreover, this study encourages the SC attributes, such as consumer evaluations and sharing on social commerce sites. Our primary focus is on community members and sellers who have provided the above-mentioned SC features. Hence, the people component is represented through the SC community members and sellers because they are the key competence of these SC features. Social commerce platforms, SA, and online payment best represent the technology. The technology and people dimensions are further discussed in sections “Institutional-Based Trust” and “Trusting Beliefs.”

Trust has been thoroughly researched and distinctly defined in several research papers. Due to the diversity of definitions for this perplexing concept, we opted to utilise a well-referenced and widely accepted definition of trust that has been cited in several pieces of research. The definition of trust by Mayer et al. (1995) has been used in this study, i.e., “a party’s willingness to be responsible for the acts of another party in exchange for the anticipation that others will perform a specific action significant to the truster, regardless of the truster’s ability to track or control the other party” (p. 712).

Trust is important in trade interactions, particularly those with unknown dangers in the Internet world (Gefen et al., 2003). The acceptance or rejection of consumers of online transactions is heavily influenced by their trust (Pavlou and Gefen, 2004). Trust is frequently interpreted as confidence in both user feedback and the capability, reliability, and honesty of retail suppliers (Hajli et al., 2017; Leong et al., 2020). Trust is a three-dimensional concept encompassing competence, integrity, and benevolence (McKnight et al., 2002). Competence means the capability of the trustee to meet the requirements of the truster; benevolence is the concern of the trustee’s for and determination to behave in the interests of the truster; integrity is defined as the capability of the trustee to keep honest commitments (McKnight et al., 2002). The trustor is the person who creates the trust, while the trustee is the one who receives it (McKnight et al., 1998).

This study responded to the appeal from previous studies (Lin et al., 2017, 2019; Hajli, 2020; Leong et al., 2020, 2021; Mou and Benyoucef, 2021; Tuncer, 2021) for more research on the factors contributing to trust in the SC context. They proposed that trust is the notion that enables customers to adopt SC after considering the features of social media and vendor suppliers, the underlying Internet infrastructure, and the payment.

Previous SC studies have recognised that trust is a significant factor that drives the behavioural intentions of customers like intention to use (Rahman et al., 2020), buying or repurchasing intent (Kim and Park, 2013; Hajli et al., 2017; Meilatinova, 2021), and aim to support (Tuncer, 2021), word of mouth (WOM) (Meilatinova, 2021), and loyalty behaviour intention (Molinillo et al., 2021). There is limited studies that emphasise SC trust and behaviour usage intention (Mou and Benyoucef, 2021; Tuncer, 2021).

Additionally, a lack of empirical investigations demonstrates the trust antecedents in SC (Lin et al., 2017, 2019; Leong et al., 2020, 2021; Meilatinova, 2021). The social factors of trust have been extensively researched (Kim and Peterson, 2017; Leong et al., 2020; Mou and Benyoucef, 2021). For instance, Leong et al. (2020) investigated trust in social commerce using networks and social presence. Their study discovered that information support had the most significant influence on the SC trust using a mixed SEM-ANN method. According to Leong et al. (2020), “the drivers of trust creation in e-commerce remain substantially unknown.” Recently, Mou and Benyoucef (2021) published meta-analysis research on customer behaviour in SC. It analyses several theoretical frameworks and examines the influence on the phases of consumer decision-making of the elements provided by them and the moderators among the variables. Interpersonal and organisational trust was found to be associated with SC behaviour. Additionally, a meta-analysis examines the factors contributing to mobile consumer engagement trust (Sarkar et al., 2020). The study found several influential elements that affect trust in mobile commerce. These elements include, but are not limited to, SA, disposition to trust, perceived security and risk, information and user interface quality, perceived usefulness, and ease of use.

In the SC literature, many trust factors have been identified. As an example, these variables include trust in SC websites and applications (Hajli, 2020), trust in sellers and social media platforms (Hajli et al., 2017; Tuncer, 2021), community trust (Cheng et al., 2019; Molinillo et al., 2021), perceived trust (Kim and Park, 2013; Leong et al., 2020), trust in mobile social commerce (Leong et al., 2021), and trust disposition (Cheng et al., 2019).

Trust was found as a one-dimensional component in most of these investigations. There is a lack of research that has seen several distinct trust variables in a single study (Lin et al., 2019; Hajli, 2020; Mou and Benyoucef, 2021; Tuncer, 2021). Additionally, to our knowledge, only one study conceptualises trust as a multidimensional component, namely, Lin et al. (2019). Their study employed the social-technical theory to conceptualise the SC trust in a multidimensional view. Social commerce trust construct included trust in e-commerce sites, trust in social media, SC features, and consumers. The analysis of data of 904 United States Amazon customers, collected by using an online survey, shows a strong provision for this conceptualisation. In addition, the results revealed that trust in social commerce features and trust in consumers have stronger impacts than trust in social media and trust in e-commerce sites in the formation of SC trust. Moreover, Lin et al. (2019) believe that “understanding the role and antecedents of customer trust in social e-commerce would give additional insight into the phenomena of social commerce and its economic effects.” Therefore, this study addresses these gaps by identifying diverse trust variables, including trust in the Internet (SA), trust in SC platforms (social networking and e-commerce sites) (multidimensional), trust in sellers (multidimensional), trust in SC community members, and cognitive trust (online payment). The next section discusses the trust constructions used in this study.

There is a significant degree of ambiguity from an SC view because of the lack of face-to-face interactions and the massive amount of user-generated material (Featherman and Hajli, 2016). As a result, building relationships with SC users is challenging as they cannot verify the reliability and authenticity of the user-generated material (Leong et al., 2020). Additionally, because the SC customers lack face-to-face contact, they cannot detect non-verbal facial movements, making trust-building in SC problematic. As a result, it is critical to understand trust in social commerce better.

Trust has long been seen as a critical component of the activity of customers in e-commerce. Trust in e-commerce is a multifaceted concept comprised of three parts: (1) trust in website, (2) trust in network operator, and (3) trust in community members (Abed, 2018; Fu et al., 2018). However, in SC notion, utilising a part of the Chinese population, it was discovered that trust in community members had no significant effect on social purchasing intent. In comparison, a poll of Netizens of the Douban.com portal discovered that trust in community members has a significant role in influencing social shopping intent (Mou and Benyoucef, 2021). This motivated us to investigate if one component of trust is more important than the others in influencing customer behaviour in social commerce, or if the effects of trust varied across different SC platforms (Mou and Benyoucef, 2021).

Figure 2 depicts the SC trust model established in this study. This model has three concept levels: (1) trusting beliefs, (2) institutional-based trust, and (3) cognitive trust. It is essential to distinguish between trust in an online service provider and faith in technology as a transaction platform (McKnight et al., 2002; Luo et al., 2010). Therefore, the model distinguishes between two sorts of trust: initial trust in the sellers, SC community members (which may be new or pre-existing trust), pre-existing trust in the SA, SC platforms, and online payment. Trust is assessed between the two distinct exchange transaction parties, as the transaction duties of the social commerce platform and sellers may vary per customer. As a result, the trust of each consumer in the SC platform and other customers may vary.

Trust in social commerce or SC trust is defined as the user’s beliefs that SC systems are reliable, and a feeling of being protected and being concerned in the online purchasing transaction (Ng, 2013; Leong et al., 2020). The phrase “trusting beliefs” refers to the opinion of a confident truster that the trustee – such as a particular Web-based vendor – possesses favourable characteristics for the truster (McKnight et al., 2002). Both trusts in sellers and in SC community members are considered to be the steadfast beliefs in this study. Trust in the sellers refers to the expectation of consumers to be able to trust the official testimony and speech of the sellers on the SC platforms (Tuncer, 2021). The term “SC community” refers to a subset of social media and online members who utilise user-generated content (UGC) and human engagement to assist customers in making product and service purchase decisions (Cheng et al., 2021). Opinions of consumers of the integrity and capability of the SC platforms to provide accurate evaluations and suggestions are described as trust in digital spaces (McKnight, 2005).

As per the model of trust-based acceptance by Komiak and Benbasat (2006), cognitive trust captures the unique trusting views of users about the trustworthy qualities of the trustee, such as professional qualifications, familiarity, and dependability. Intellectual trust will be created to a high degree if people find adequate causes to trust (Gong et al., 2020). Cognitive trust in online purchases is used in this study to refer to the belief of users in the potential of SC platforms to offer accurate and trustworthy financial services (Gong et al., 2020; Leong et al., 2021). It measures the cognitive views of consumers about the dependability and credibility of online payment systems. Trust has been recognised as a critical element in the research of online human behaviour, particularly in transactions like SC, where customers are more receptive to online payments (Williams, 2021).

Trust in the Internet and SC platforms is the pre-existing trust based on institutions (Hajli, 2020; Tuncer, 2021). Institutional trust is defined as “the assumption that necessary structural circumstances exist (e.g., on the Internet) to increase the chance of success in an endeavour such as e-commerce” (McKnight et al., 2002, p. 339). Trust in institutions (through the Internet) is described as having a single dimension: SA. SA is “belief in the existence of structures like as guarantees, rules, pledges, legal remedies, or other mechanisms designed to encourage success” (McKnight et al., 2002, p. 339). A customer with a high level of structural certainty regarding the Internet would think that technical and legal safeguards, such as data security, would protect him/her from financial, privacy, or identity information loss. The SC users feel that SC sites on the Internet are competent and trustworthy. The term “trust in the social commerce platform” relates to the views of consumers of the institutional structure of social media and e-commerce platforms system, and their sentiments about the SA of the system (Pavlou and Gefen, 2004; Tuncer, 2021).

The suggested model analyses the relation between trust in the SC platform and the Internet. According to cognitive dissonance theory (Bhattacherjee and Premkumar, 2004), individuals may be less inclined to provide their details if the environment that the Internet provides is considered unsafe, irrespective of the level of trust they have in a specific supplier. On SC platforms, the lack of face-to-face interaction generates confusion and insecurity (Featherman and Hajli, 2016). According to one study, trust in the Internet and e-commerce businesses is critical for making online purchases (Kim and Peterson, 2017).

Prior IT adoption research discovered a tangible link between SA and trust (Luo et al., 2010; Zhou, 2012; Oliveira et al., 2014; Khan et al., 2021). Moreover, mobile social commerce studies confirmed the substantial relationship between SA and SC trust (Sarkar et al., 2020). A significant relationship between social media interaction, SC applications and websites, and institution-based trust was discovered by Hajli (2020), leading to purchasing intentions on SC websites. Furthermore, SC trust is influenced by e-commerce and social networking sites (Lin et al., 2019). Likewise, SC users with a positive impression of SA and the SC platform are more inclined to share their data and utilise the SC. As a result, we predict that trust in the SC platform and SA influence the willingness of consumers to trust in SC. Accordingly, we posit the following hypotheses:

Prior trust studies have shown that SA of the platform of the context has a more significant effect on trusting views (McKnight et al., 2002; Li et al., 2008; Khan et al., 2021). Moreover, SC consumers are more likely to have favourable, trusting attitudes in social commerce if they feel that sellers, SC communities, and online payment service providers on the Internet are run with truthfulness, sincerity, and compassion. In the context of technology, technological protections like encryption, particular system development methods, and techniques have been discovered to assist early trusting beliefs in an online seller and website (Li et al., 2008; Luo et al., 2010). Social commerce customers may also look for Internet rules and regulations and legal ways to secure their information (Jiang et al., 2021). We argue that trusting beliefs in online sellers, community members, and payment will enhance due to the SA connected to the Internet in general. As a result, we anticipate that users will be more likely to trust social commerce systems if they feel comfortable engaging with the Internet in general. As a result, the following hypotheses emerge:

The safety, security, and guarantees provided by the e-commerce site or place of business of the seller form the basis of trust (Hajli, 2020). When customers trust an SC platform, they are more likely to purchase and engage in social interactions (Hajli, 2020; Tuncer, 2021). Individuals use SC platforms and make purchases when they feel the platforms are trustworthy and satisfy their needs (Rahman et al., 2020). Furthermore, the willingness of people to adopt social commerce platforms increases when their trust in the web vendors is positively influenced (Lu et al., 2016). Hence, we postulate that:

Social commerce platform features such as reviews, forums, ratings, and communities are critical predictors of customer behaviour in the SC sector (Hajli and Sims, 2015; Cheng et al., 2021). Trust and perceived risks can be improved by online communities (Hajli, 2015). Trust can be maximised, and risks connected with online transactions can be minimised through interactions via the members and features of the online community on social commerce platforms and Internet technologies (Hajli, 2013; Rahman et al., 2020). Thus, we hypothesise:

Customers can make purchases employing online payment methods without revealing their personal information like credit card details by using several SC applications and websites available today (Mou and Benyoucef, 2021; Williams, 2021). Although Line, Facebook, Instagram, and other sites are primarily used for social networking, numerous commerce activities, like purchases, are also possible. Customer trust in the vendor as a prerequisite to purchasing decisions, evaluations, and shopping is becoming more critical as the market is brimming with purchase options (Hajli, 2020). Hence, customers are more likely to trust social commerce vendors and conduct online financial transactions if they have more vital institution-based trust in social commerce platforms.

Regarding the expertise, knowledge, and experience of social commerce sellers, buyers may be concerned about their competence and whether they are working for themselves or on behalf of someone else to meet their demands (Wongkitrungrueng and Assarut, 2020). By providing assistance in information and in notifying consumers that they care about their requirements, the generosity of sellers may be induced. Thus, they may promote their credibility by making good faith claims about the quality of the business offerings (items and services) they deliver to their consumers. Trust can assist SC consumers in overcoming their worries and encouraging them to use social commerce. Some businesses, such as Walmart, have failed to implement SC because of numerous security, privacy, and trust concerns (Kim and Park, 2013). Customers refrain from making e-commerce purchases by trusting the sincerity, competency, trustworthiness, and customer feedback of sellers (Hajli et al., 2017).

Trust is a challenging phenomenon in social commerce compared to traditional e-commerce (Gefen et al., 2003; Lin et al., 2019). Users interact with individual businesses under the social commerce C2C paradigm. The identities on both sides of the interaction are opaque, and there is a significant information imbalance (Leong et al., 2021). For example, sellers could make Facebook fan pages for managers to communicate directly with customers to enhance their relationships (Leong et al., 2021). Furthermore, seller and buyer interactions can be enabled through chat plug-ins in websites or chat applications, such as QQ, WeChat, etc. This can help customers receive more information and know the attitude, trustworthiness, and compassion of sellers to build their trust beliefs (Leong et al., 2020). Such factors make customers believe that purchasing decisions are risky where trust becomes visible and a severe issue (Tuncer, 2021). Hence, the ease and assurance required to make purchases through SC platforms rely on the level of trust of the seller (Wongkitrungrueng and Assarut, 2020). We propose the following hypothesis based on our analysis:

As a new e-commerce medium, social commerce incorporates social aspects and fosters consumer-generated material, such as suggestions, comments, images, reviews, and online streaming (Hajli and Sims, 2015; Cheng et al., 2021). Before making a purchasing decision, many customers seek the advice and suggestions of others (Chen et al., 2017). Social commerce offers significant benefits, such as linking merchants and consumers, and assisting them in acquiring and sharing product knowledge (Xiao et al., 2015).

In the context of interpersonal communication, social commerce networks may provide customers with a very engaging environment where they can trust and connect while also expressing their views on a product or service (Liang et al., 2011; Cheng et al., 2021). Trust in members is influenced by the social advantages of the community that customers acquire (Tsai and Hung, 2019). When consumers receive trustworthy, accurate, and timely information and tools from SC communities, their trust increases substantially (Yeon et al., 2019; Cheng et al., 2021). For a good shopping experience and better purchasing decisions, customers in SC depend on informative support and product reviews (Yang et al., 2021). According to Leong et al. (2021), online reviews significantly affect trust in mobile social commerce. Furthermore, Lin et al. (2019) discovered a strong relationship between social media elements (i.e., product ratings, reviews, and sharing) and social commerce trust. Hence, we hypothesise:

Online payment is critical to completing social commerce operations (Zhang and Benyoucef, 2016; Williams, 2021). Assumptions of customers that the SC online payment service features are credible are referred to as cognitive trust in the online payment (Leong et al., 2021). Previously, cognitive trust was defined as trusting attitudes involving trust in mobile and online payments (Gong et al., 2020; Leong et al., 2021). Trust in mobile social commerce is significantly influenced by cognitive trust in mobile payments (Leong et al., 2021). According to Lin et al. (2011), trust can be transmitted intra-channel when the trust of the customers is transferred to another entity within the shared channel. As both online payment and SC are part of the same online service-to-service intra-channel, we predict that the trust of customers in online payment will be extended to trust in social commerce services. Thus, the following is our hypothesis:

The extent to which an individual considers whether to indulge in a specific activity or the likelihood of the intention of an individual to perform a specific behaviour is known as behavioural intention (Venkatesh et al., 2013). This paper describes behavioural intention as the probability that people will use a social commerce site in the future to buy online. Customers who build trust in a particular SC platform are more likely to return (Meilatinova, 2021). According to Ng (2013), the intention of a customer to buy from a social commerce site is enhanced through trust in an SC platform. Moreover, a significant relationship exists between buying behaviour and social commerce trust. According to the results of a meta-analysis of SC research (Sarkar et al., 2020; Mou and Benyoucef, 2021), there is a substantial relationship between trust and customer behaviour intention.

As a multifaceted concept, social commerce trust impacts customer decisions as a holistic process that combines the cognitive and emotive processes by incorporating the effects of its aspects (McKnight et al., 2002; Lin et al., 2019). Customers with a high level of social commerce trust believe that SC platforms, online payment systems, and Internet technology are reliable and satisfy their needs during the online purchasing process. Consumers will also believe that vendors and SC communities are trustworthy and concerned about their requirements, leading to favourable impressions of trust in SC and a desire to continue using it. As a result, we propose the following hypotheses:

As control variables, this study includes demographic factors such as age, gender, and education. Prior empirical study has shown that individual user variations in behavioural intentions toward technology adoption are influenced by demographic factors (Venkatesh et al., 2012, 2013). Moreover, researchers have indicated that these variables are crucial to consider in customer surveys, emphasising the need to control them in uncertain behaviour, such as trust in social commerce (Booth and Nolen, 2012; Featherman and Hajli, 2016; Leong et al., 2020). Gender and age were found to influence trust in information technology by Pak et al. (2014). Furthermore, trust is significantly affected by education and gender (Calabrese et al., 2016). Based on prior findings, it seemed appropriate to include gender, age, and education as control variables in this study.

This study uses a target sample of individuals who were older than 18 and had online shopping experiences through the SC platforms (e-commerce and SC websites or social networks platforms). Social networks sites (SNS), specifically, Facebook and Instagram users, were the target population. There are several reasons to choose this sample set. First, SNS is very popular, with millions of actual and potential SC users (Mou and Benyoucef, 2021). Second, Facebook and Instagram are among the top three SC providers with various online community and payment features and the highest SC users (Mou and Benyoucef, 2021). Third, different technical mechanisms and functionalities are conducted on both platforms. This is because they are attractive social networks, and it is crucial to investigate the key socio-technical differences. Furthermore, these socio-technical differences are significant to examine changes in beliefs and perceptions among responders (Dwyer et al., 2007), which enabled the demographic differences to construct the sample. Finally, research (Sibona and Walczak, 2012) reported that recruitment through SNS is suitable to gain insights into the behaviour of consumers and helpful in understanding the technology adoption. However, most previous SC adoption studies have focussed on Facebook as an SNS platform, and limited research has studied trust in social commerce from dual SNS platforms or different sites such as Instagram (Leong et al., 2020; Tuncer, 2021).

Facebook and Instagram are commonly used in almost every country in various languages. Moreover, data collection in different countries and cultures is challenging due to the language and cultural differences (Van de Vijver and Chasiotis, 2010). Therefore, this study is conducted in three English-speaking countries, namely, (i) Australia, (ii) the United States, and (iii) the United Kingdom. The selection of these countries is based on certain reasons. First, the survey was designed in English to ensure that the respondents can understand the questionnaire. Second, data collection from diverse settings could help to increase the generalisability of the research findings (Van de Vijver and Chasiotis, 2010). Third, Facebook and Instagram have been reported as the most popular social networks, with more users across the selected countries (Statista, 2021a,b).

Practically, collecting the information of Facebook and Instagram users is crucial. However, these platforms enable individuals to target users via advertising. Therefore, this study employs a banner advertisement placed through Facebook Adv and Instagram Adv as a technique to recruit subjects (Alvarez et al., 2003). This technique enables advertisers to determine which members view the ads by selecting target viewers based on age, gender, location, and interests. This study ran two banner campaigns through each site for 2 weeks, from 15 March 2021 to 30 March 2021.

In total, 843 people clicked on the advertisement and were directed to the survey website. The initial data analysis revealed that 258 (30.6% of the individuals directed to the survey website) did not complete the questionnaire and exited the survey at some point before completion, and only 585 individuals completed the questionnaire. Furthermore, knowing whether an individual has a prior online shopping experience with SC platforms, the filtering question of “Have you been involved in social commerce (e.g., buying online through social media) before?” was asked (Leong et al., 2020). Among 585 completed responses, some data were discarded. For example, 112 responses were ruled out since the participants did not have a prior online shopping experience with SC platforms. In addition, 67 individual responses were removed because they were invalid, either having the same answer to most questions or completing the survey in less than the average time (7 min). Therefore, 406 responses (usable data sets) were used for further analysis. The yielded sample size was adequate as it fulfilled the criteria for conducting confirmatory factor analysis using PLS (Goodhue et al., 2012; Hair et al., 2019).

The measurement items were adopted from previous well-known researches to sustain construct and content validity (see Supplementary Appendix A). Gefen et al. (2011) argued that the total of three items loading on one construct should statistically identify the factor measurement. Accordingly, all constructs were measured by at least three items except for behavioural intention. The behavioural intention is defined as the extent to which individuals will continue purchasing using a particular SC platform in the future and is measured with five reflective items that have been modified and adapted from Bansal et al. (2010); Venkatesh et al. (2012), and Aren et al. (2013). Trust in SC was adapted from Kim and Park (2013). SA was adapted from McKnight et al. (2002). Trust in SC platform and trust in Sellers constructs (benevolence, competence, and integrity) were adapted from McKnight et al. (2002); Li et al. (2008), and Wongkitrungrueng and Assarut (2020). Trust in the SC community construct was adapted from Cheung et al. (2009); Hajli (2015), and Rahman et al. (2020), and trust in online payments items were adapted from Leong et al. (2021). A seven-point Likert scale that ranged from “strongly agree” to “strongly disagree” was used to measure the scales (Hinkin, 1998). A pre-test phase was considered to refine and validate the questionnaire (MacKenzie et al., 2011). The draft of questions and measurements was sent for review to 10 social networkers, who are online shoppers, and five information systems (IS) professors, seeking their confirmation of the face validity of the survey questionnaire. Based on their feedback, minor changes and revision were made.

This study implemented some methods to assess and mitigate the influence of common method bias (CMB), as recommended by Barlette et al. (2021). A priori procedural remedy by Podsakoff et al. (2012) was included. This method was used during the pre-test phase to refine the scale items and to eliminate potential ambiguities, with multiple-choice questions periodically included to break up the pattern of questions rated using Likert scales. In addition, Harman’s single-factor test was conducted. The results of the Harman one-factor test showed that the total variance was less than 50%, indicating that CMB was not a concern in this study (Podsakoff et al., 2003). Furthermore, the path coefficients and correlation of the constructs obtained from the structural model assessment had varying degrees of significance (Barlette et al., 2021; Williams, 2021). These results suggested that CMB was not a problem in data.

In addition, a non-response bias test was conducted (Churchill, 1979). Two groups were formed according to the order in which participants completed the survey questionnaire. The first sub-sample consisted of the first 100 early respondents, while the second included the last 100 late respondents (Churchill, 1979). A two-tailed t-test at a 5% significance level was assessed to compare these two sub-sample groups (Field, 2009). The test findings showed no significant differences between these two groups of respondents. Hence, non-response bias was not a problem in this study.

Table 1 shows the demographic profile of respondents. The sample is majorly comprised of male respondents (55.6%). Moreover, in the sample, most of them were young, and the ages ranged from 18–41 (86.6%), with the remaining 13.4% older than 42 years. The subjects were relatively highly educated (85.5% had a Bachelor’s degree or higher). Concerning their online shopping history, 40% of the sample had frequently shopped online. The majority of respondents were mainly from the United States, Australia, and the United Kingdom, respectively.

This study employs PLS using SmartPLS 3.0 software (Ringle et al., 2015) to analyse the data and to assess the structural relationships of the research model. This study employs the confirmatory composite analysis (CCA) approach, which is recently introduced and applied to confirm the measurement models when using PLS-SEM (Hair et al., 2020; Hair, 2021). This study follows the guidelines of Hair et al. (2020) in accessing PLS- CCA to confirm the reflective measurement model. Moreover, this study uses the PLS technique because it does not build a strict assumption about the normal data distribution and is preferable for composite analysis (Hair et al., 2021). In addition, non-parametric bootstrapping was carried out using 406 cases and 10,000 samples (Streukens and Leroi-Werelds, 2016) to obtain the significance of each structural path (i.e., the t-value) between the constructs (Hair et al., 2021).

This study is based on a structural model that contains six unidimensional (single) constructs (i.e., behavioural intention, SC trust, trust in SC community members, trust in online payment, and SA), two hierarchical (second-order) constructs (i.e., trust in SC platform and trust in sellers). Previous studies on trust in online contexts (McKnight et al., 2002; Li et al., 2008; Luo et al., 2010; Alkhalifah and D’Ambra, 2013) and social commerce (Lin et al., 2019) have specified these variables as a hierarchical construct consisting of few sub-dimensions and dimension in most cases.

The results of the path analysis and hypothesis testing are presented in Table 6. The results show that among the 13 directly hypothesised relationships in the research model, two were non-significant.

Our findings suggest that the institution-based trust, SA (β = 0.031), has no influence on SC trust, whereas trust in SC platform (β = 0.255, p < 0.01) has a positive impact on SC trust. Therefore, H1 was not supported while H3 was supported. In addition, the proposed influence of SA on the trust in SC platform (H2) (β = 0.747, p < 0.001) was significant. SA was positively related to trust in sellers (β = 0.132, p < 0.001), trust in SC community members (β = 0.297, p < 0.001), and online payment (β = 0.198, p < 0.01) (H4, H5, and H6, respectively). Moreover, H7, H8, and H9 were all supported with the positive relationships found to be significant between trust in SC platform and trust in sellers (β = 0.822, p < 0.001), trust in community members (β = 0.525, p < 0.001), and online payment (β = 0.642, p < 0.001). The positive associations between trust in SC community (β = 0.191, p < 0.01) and online payment (β = 0.256, p < 0.001), and SC trust were significant, whereas the relationships between trust in sellers and SC trust (β = 0.16) were not significant. Thus H11 and H12 were supported, whereas H10 was not accepted. Finally, the hypothesised relationships related to the effect of SC trust on behavioural intention (H13) (β = 0.628, p < 0.001) were supported.

This study relied on Explained Variance (R2) to confirm whether the model achieved acceptable goodness of fit as illustrated; there was no other overall parametric criterion in PLS (Hair et al., 2019). The overall variance explained by the research model in terms of R2 was 0.463 for behavioural intention, 0.597 for SC trust, 0.638 for online payment, 0.637 for trust in SC community, 0.612 for trust in SC platform, and 0.849 for trust in sellers, as shown in Table 8.

This study also adopts Cohen’s (1988) effect size (f2) test for first-order constructs in the model. The sizes of effects are small, medium, and large with f2 values of 0.02, 0.15, and 0.35, respectively (Cohen, 1988; Sarstedt et al., 2017; Hair et al., 2019). The effect of size for SC trust on behavioural intention was high (0.702). Also, it was a large (1.34) adequate size for SA on trust in the SC platform. The effect size on the endogenous variable SC trust was highest for trust in the online payment (0.060), lowest for trust in the SC platform (0.26), and both indicated medium effects. The f2 effect size on the trust in seller was largest for trust in SC platform (1.73) and lowest for SA (0.046), indicating high and medium effects, respectively. Similarly, the effect size on the online payment was largest for trust in the SC platform (0.442) and lowest for SA (0.043), indicating high and medium effects, respectively. Moreover, the f2 effect size on the trust in the SC community was largest for trust in the SC platform (0.296) and lowest for SA (0.098), indicating medium and small effects, respectively.

This study evaluated further testation in this regard and rigorously determined the implications of predictive relevance (Q2). The predictive relevance, or the predictive sample reuse technique, is vital for testing the predictive validity of a complex model. It also helps measure how well-examined values are replicated by the model (Sarstedt et al., 2017). A blindfolding procedure was used with the omission distance of 7 to calculate Q2. This study obtains a cross-validated redundancy Q2 of all the focal variables, as shown in Table 8. The result of Q2 values exceeded zero (Q2 > 0), which revealed a highly predictive model (Sarstedt et al., 2017; Hair et al., 2021).

Furthermore, the PLS predicted algorithm was assessed to measure the predictive power (Shmueli et al., 2019). The predictive significance of the model was determined by comparing the mean absolute error (MAE), root mean square error (RMSE), and Q2 predict values for the PLS-SEM model against the naive benchmark model (LM) (Hair et al., 2019; Shmueli et al., 2019). Most indicators of the endogenous variables had positive Q2 values except BI1, COPT3, TSA1, TSCPC3, and TSELLC3. Moreover, the majority of indicators scored lower on the MAE in the PLS-SEM than the LM as shown in Supplementary Appendix B. This indicated that the structural model of this study shows a medium predictive power.

This study has linked the control variables to all the endogenous variables in the model as recommended by Atinc et al. (2012) and Carlson and Wu (2012), with testing the impact of control variables. We assumed the path coefficient and t-value for endogenous variables, as shown in Table 9. The results yielded that five of 18 relationships (3 covariates* 6 dependent variables) were significant.

This study has a number of limitations that could be further addressed in future research. First, trust in the SC platform was conceptualised and was tested through the limited dimensions of SC platforms (e-commerce and social media sites). This study had a shortcoming of differentiating trust issues in various SC platforms. Different SC platforms, such as e-commerce websites, social media platforms, and SC mobile applications, could be tested separately to investigate their effects on SC trust and behavioural intention. Second, trust in sellers was not found to be significant on trust in SC. This possibly reflected the limited assessment of this factor which was only evaluated concerning the integrity and reliability of the sellers. Incorporating and examining more factors toward trust in sellers, such as information technology affordance (Tuncer, 2021) and symbolic value (Wongkitrungrueng and Assarut, 2020), could enhance trust in sellers leading to the reliability of SC platforms. This is how purchasing can be made dynamic. Third, the selected subjects were limited to only three English-speaking countries. Therefore, future research could evaluate the model in different settings and countries to confirm the generalisability of the findings as cultural differences exist among countries, which lead to different perceptions and behaviours (Venkatesh et al., 2012). Fourth, only trust in SC was measures concerning behaviour intention, which yielded 46% of explained variance. Future studies may test the moderation effects and incorporate other variables such as social factors, characteristics of individuals (e.g., interpersonal trust), privacy concerns, and design features into the model. After testing their direct, mediation, and moderation effects on both behavioural intention and trust in SC, new doors to the success of the SC industry can be brought.

Finally, this study evaluated and confirmed the best fit of the research model using a single SEM-PLS analysis method. Future studies may confirm the model validity and predict the performance of the variables using a dual-approach analysis employing machine learning tools such as artificial neural networks (ANN). Dual-approach (PLS_ANN) analysis has been recently suggested in social commerce (Leong et al., 2020) and information technology adoption research (Alwabel and Zeng, 2021).

Due to the increased number of social commerce users, existing researches have been conducted to consider the consumer behaviour on different SC websites. However, research paid little attention to the investigation of multi-facets trust attributes in social commerce. Hence, this study developed a research model based on social-technical theory and identified different trust facets to examine the importance of trust dimensions in social commerce. Using online surveys and data (N = 406) from three different countries, the empirical results contribute to theory and practice by explaining the relationships identified in the model. The findings of this study enhance our understanding of SC trust and open new debate in social commerce research.