Need a Theory Factsheet about a case study, the lentgh of work 2 pages of writing.

Social commerce components Social learning theory Forums and communities Ratings and reviews Social recommendations ABSTRACT The proliferation of social commerce has changed customers ’purchase decision-making process. However, few studies have investigated the roles of social commerce factors on customers ’purchase decision-making. Based on the social learning theory, we develop a research model to examine how customers ’learning behavior along three main social commerce components (SCCs) a ffects customers ’attitude in both cognitive and a ffective di- mensions and how such attitude determines customers ’purchase intention. The results from a survey of 243 actual users of social commerce websites suggest that cognitive and a ffective appraisals are the main predictors of purchase intention, with cognitive appraisal having a higher predictive power than a ffective appraisal. In addition, learning from forums and communities and learning from ratings and reviews have signi ficant infl u- ences on both cognitive and a ffective appraisals, while learning from forums and communities plays a more important role in formulating a ffective appraisal and learning from ratings and reviews plays a more important role in determining cognitive appraisal. Contrary to our expectation, learning from social recommendations has no signi ficant infl uence on either cognitive or a ffective appraisal. In summary, these findings provide a com- prehensive understanding about customers ’purchase decision-making process and extend the application scope of social learning theory. The findings also provide social commerce managers guidance in designing more eff ective websites and allocating resources and eff orts reasonably on different SCCs. 1. Introduction Social commerce is a form of commerce mediated by social media involving convergence between the online and o ffline environments ( Bai, Yao, & Dou, 2015 ;Chen & Shen, 2015; Shanmugam, Sun, Amidi, Khani, & Khani, 2016). Social commerce websites provide various ap- plications including product recommendation to a friend, customer review provision, discussion board, and writing and rating a review, all of which are called social commerce components (SCCs) ( Hajli, 2015).

Within this environment, customers have access to social knowledge and experiences that support them to better understand their online purchase purpose and make more informed and accurate purchase decisions ( Huang & Benyoucef, 2015 ). However, how SCCs facilitate customers ’purchase decision is not yet fully understood by both aca- demia and industry. From the corporate perspective, quickly and e fficiently im- plementing the social commerce business model is both rewarding and challenging. Hajli (2015) points out that adding SCCs in existing busi- ness platform is the most e fficient and e ffective way to implement social commerce. Since resources are limited, managers are looking for a quick start in developing the speci fic SCC which is key to facilitate customers ’purchase, other than distributing resources on all the SCCs equally ( Chen, Lu, & Sumeet, 2017 ). Thus, understanding the di fferent roles of various SCCs in customers ’decision-making process can provide guidance in allocating resources in di fferent SCCs. On another hand, the primary revenues of a majority of social commerce websites only rely on advertising ( Kim, Gupta, & Koh, 2011). A unitary pro fit pattern would make the company face a high risk when the market changes. By having a better understanding of the roles of SCCs on customers ’pur- chase intention, social commerce websites can implement these stra- tegies to obtain alternative or additional sources of revenue. From the customer perspective, customers gather information in the process of purchasing a product to aid their decision ( Bai et al., 2015).

In e-commerce, studies have shown the factors surrounding the mer- chant (e.g., website design, reputation, service quality) or product (e.g., speci fication, quality) or individual factor (e.g., trust, self-e fficient) signi ficantly impact the consumer purchase decision ( Bai et al., 2015; Li, Wu, & Lai, 2013 ;Luo, Ba, & Zhang, 2012). In social commerce, http://dx.doi.org/10.1016/j.ijinfomgt.2017.05.001 Received 15 September 2016; Received in revised form 5 May 2017; Accepted 5 May 2017 ⁎Corresponding author.

E-mail addresses: [email protected] (A. Chen),[email protected] (Y. Lu),[email protected] (B. Wang). International Journal of Information Management 37 (2017) 627–638 0268-4012/ © 2017 Elsevier Ltd. All rights reserved. MARK customers make their purchase decisions by participating in online social media to acquire social knowledge about the product they want (Huang & Benyoucef, 2015 ). It is also possible that the purchase sti- mulus is aroused by browsing social media ( Chen et al., 2017; Kim & Park, 2013 ). Customers may also look at comments from other customers and social recommendations on social commerce sites ( Zheng, Zhu, & Lin, 2013) or solicit opinions from their friends ( Lecinski, 2012; Shin, 2013 ). Before customers visit the store shelf, they have already made their decisions. However, there have been very few studies discussing the e ffect of social information on the consumer purchase decision. Besides, both online and o ffline buyers are increasingly using the social media as a source of information to assist them in purchasing ( Rondán-Cataluña, Arenas-Gaitán, & Ramírez-Correa, 2015 ). The greater accessibility to di fferent information sources that provides the social media not only might help consumers to be less susceptible to potential deceptive o fferings online, but also might be critical in driving their perceptions and behaviors when they shop o ffline ( Riquelme & Román, 2014 ). The cross-channel effects (i.e., get in- formation online, but purchase o ffline, or vice versa) are also sig- ni ficant, which has been indicated by recent research ( Dinner, Van Heerde, & Neslin, 2014 ). Given that consumers use more online media when making online purchases than when making o ffline purchases and consumers use o ffline media to the same extent when buying online or o ffl ine ( Voorveld, Smit, Neijens, Segijn, & Bronner, 2014 ), this study investigates how online SCCs facilitate customers ’online purchase de- cision. In this study, we incorporate three main SCCs (forums and com- munities, ratings and reviews, and social recommendations) in the re- search model. Based on the social learning theory ( Bandura, 1978), we hypothesize that learning from the three SCCs can in fluence customers ’ cognitive appraisal and a ffective appraisal during their purchase deci- sion-making process, which in turn in fluences their purchase intention.

Findings of this study enrich the existing literature on social commerce and deepen our understanding about the di fferent roles of various SCCs in changing customers ’purchase decision-making process. In addition, our results provide social commerce operators practical implications on how they should design their websites and allocate resources in various SCCs in a more optimal fashion.

The rest of the paper is organized as follows. The second section presents the theoretical background and literature review. After that, we discuss hypotheses development, the research methodology, and the results. Finally, we conclude with a discussion of the results, the con- tributions and limitations of the study, as well as avenues for future research.

2. Theoretical background 2.1. E-commerce and social commerce The di ff erences between e-commerce and social commerce can be highlighted in terms of business goals, customer connection and system interaction ( Huang & Benyoucef, 2013, 2015). With regard to business goals, e-commerce focuses on maximizing e fficiency with strategies for sophisticated searches, one-click buying, speci fication-driven virtual catalogs and recommendations based on consumers ’past shopping behavior. Social commerce, however, is oriented toward social goals, such as networking, collaborating and information sharing, with a secondary focus on shopping ( Wang & Zhang, 2012). Regarding cus- tomer connection, customers usually interact with e-commerce plat- forms individually and independently from other customers, while so- cial commerce involves online communities that support social connection to enhance conversation between customers ( Zhang & Benyoucef, 2016 ). As for system interaction, e-commerce in its classical form almost always provides one-way browsing, where information from customers is rarely sent back to businesses or other customers. Social commerce, however, develops more social and in- teractive approaches that let customers express themselves and share their information with other customers as well as with businesses ( Shanmugam et al., 2016).

2.2. Social commerce components Yadav et al. (2013) defines social commerce as the “exchange-re- lated activities that occur in, or are in fluenced by, an individual's social network in computer-mediated social environments, where the activ- ities correspond to the need recognition, pre-purchase, purchase, and post-purchase stages of a focal exchange” . Recent research identified two major types of social commerce: (a) traditional e-commerce web- sites that add social tools to facilitate social interaction and sharing; and (b) social networking sites that incorporate commercial features to allow transactions and advertisements ( Zhang & Benyoucef, 2016; Zhou, Zhang, & Zimmermann, 2013). That means there are two essen- tial components (i.e., commercial components and social components) no matter which type the social commerce service is. The commercial components have been investigated largely, such as website design, reputation, service quality, product, trust and self-e fficient ( Bai et al., 2015; Li et al., 2013; Luo et al., 2012 ). Then what do the social com- ponents incorporate and how do they in fluence customers ’decision?

For many customers, shopping is a social experience, and they often want to get other people ’s opinions before making a purchase. Contrary to the traditional e-commerce, social commerce introduces a set of so- cial components such as online forums, consumer ratings and reviews, and social recommendations to e-commerce ( Li et al., 2013; Rad & Benyoucef, 2010 ). The social commerce components/constructs (SCCs) is defi ned as the constructs that are derived through social commerce such as online forums, ratings, communities, reviews and recommendations ( Chen et al., 2017; Hajli, 2015; Shanmugam et al., 2016 ). The information produced by these social commerce platforms or communities can a ffect consumers ’purchasing intention or behavior ( Shanmugam et al., 2016). Companies use these SCCs as a platform to communicate with customers and to enable customers to communicate with each other, leading to a new channel for customer relationship management ( Hajli, 2012b; Liang & Turban, 2011 ). Social commerce enables businesses to build customer trust, increase sales, and decrease marketing cost. It brings new challenges and advantages to the online shopping experience, triggering the analysis of consumer buying be- havior in social commerce ( Rad & Benyoucef, 2010). Although some other types of components such as social advertising and social shop- ping can be included in SCCs, most researchers identify forums and communities, rating and reviews, and social recommendations as the most important social components in changing customers ’decision- making process ( Chen et al., 2017; Hajli, 2012b; Huang & Benyoucef, 2013; Kim & Park, 2013; Marsden, 2010; Rad & Benyoucef, 2010 ).

Speci fically, forums and communities are useful and e ffective social media tools for social commerce that assist product discovery, selection and referrals by providing a moderated environment around a parti- cular theme, task or category ( Shadkam & O'Hara, 2013). Forums and communities can be used as a source of product know-how and help retailers to provide user-generated content that can engage customers and drive sale ( Hajli, Lin, Featherman, & Wang, 2014 ). It is also argued that consumer online communication in social commerce constructs provides social support, which leads in turn to trust in the network ( Chen, Lu, Wang, Zhao, & Li, 2013 ). These types of support, which can be both informational and emotional, persuade individuals to reuse the system or try a new product again. Hence, these communities are va- luable sources of innovation for the marketplace. In addition, commu- nities are the main drivers of change from e-commerce to social com- merce ( Huang & Benyoucef, 2013 ).

Ratings and reviews are original social commerce toolset that allow people to exchange product feedback and inform each other ’s choices with independent views and experiences ( Shadkam & O'Hara, 2013). A. Chen et al. International Journal of Information Management 37 (2017) 627–638 628 From business perspective ratings and reviews platforms help retailers integrate customer feedback and community features directly into their websites. On the other hand, from consumer perspective the word from the independent users and real consumers on social media is more credible and useful for purchasing decision (Kim & Park, 2013).

Social Recommendation refers to using social media to get and make recommendations on what to buy, read, eat, see and do ( Shadkam & O'Hara, 2013 ). By combining the recommendation func- tion with social network applications, social recommendations assist users in making choices among various alternatives ( Chang & Hsiao, 2013 ). On social commerce sites, social recommendations are made by other customers, who have the actual purchase experience. Then social commerce sites make these recommendations available to other cus- tomers who share similarities with the recommenders in socialization, preference, and social relations ( Li et al., 2013; Marsden, 2010).

There are three signi ficant di fferences among these three SCCs.

First, the information in forums and communities is much richer and more vivid than that in ratings, reviews and social recommendations.

Customers can share their knowledge or experience in details using text, audio, video or pictures. The information in ratings are numerical ratings, typically ranging from one to five stars –from very low (one) to very high ( five) ( Hajli et al., 2014 ). The reviews usually are comprised by a piece of short text and a few pictures. And a speci fic social re- commendation is usually a picture combining a piece of information which persuades other customers to buy the products. Second, the level of information relevance among the three SCCs is di fferent to customers. Forums and communities provide a moderated environment around a particular theme, task or category, while ratings, reviews and social recommendations are produced against one speci fic product or seller. Whereas ratings and reviews are generally visible to all, social recommendations are usually personalized online goods and services designed to realize the referral value of customers and ad- vocates ( Shadkam & O'Hara, 2013 ). Customers obtain differently re- levant information in these three SCCs to aid their purchase decision. Third, customers ’purposes of browsing these three SCCs are dif- ferent. Customers participant in forums and communities based on their interests or target goods. They obtain information support and emo- tional support in forums and communities. Ratings and reviews are browsed only when customers know what they want to buy. They just evaluate the quality of the speci fic product and seller using the ratings and reviews. Social recommendations refer to get and make re- commendations on what to buy, read, eat, see and do. Social commerce sites make these recommendations available to other customers who share similarities with the recommenders in socialization, preference, and social relations.

In our proposed model, we combine the above-mentioned SCCs together to evaluate their e ffects on social shopping decision-making from the social learning perspective.

2.3. Social learning theory Information search during the shopping process is similar to window shopping, which transfers information related to a product from the vendor ’s website to a consumer ( Pavlou & Fygenson, 2006). In social commerce, customers get social knowledge about the products they are interested in through browsing the contents generated by others ( Huang & Benyoucef, 2013 ). This is a social learning process.

Social learning theory has its roots in the work of psychologists Bandura and McClelland (1977) .Itoff ers a structured approach to dealing with a variety of behavioral concerns in a number of di fferent disciplines and settings. Its use in the social and behavioral sciences as a mental health intervention has grown with waning interest in the in- sight-oriented approaches ( Chavis, 2012). It is also an in fluential theory of learning and human development and is rooted in a number of basic concepts of traditional learning ( Bandura & McClelland, 1977). How- ever, the theory adds a social element. It underlines that people can learn new information and behaviors by observing other people. Thus, the use of observational learning, imitation, or modeling explains a wide variety of human behavior using social learning theory and ap- proach.

Social learning is fundamentally learning by observing the actions of other people ( Lorenzo, Kawalek, & Ramdani, 2012). Individuals learn which behaviors are acceptable and/or unacceptable by observing and imitating others. Imitation is a function of successful modeling. Via a successful modeling process, an individual is able to quickly reproduce the behavior exhibited by the model ( O’Fallon & Butter field, 2012).

Modeling can therefore be viewed as a means of dispersing the social context ’s values, attitudes, and behaviors and is e ffective when in- dividuals are able to apply the modeled behavior to a situation ( Weaver, Treviño, & Agle, 2005).

In social commerce, the essential social psychology is social learning: learning from the knowledge and experience of others we know and/or trust ( Marsden, 2010). SCCs provide the environment for customers to observe the behavior of others. Customers have access to social knowledge and experiences by interacting with online forums and communities, browsing ratings and reviews posted by others, or considering recommendations from social networks ( Huang & Benyoucef, 2013). In this process, they become clear on what they actually want to buy, whether the products satisfy their needs, whether the sellers are reliable, and whether the shopping experience is enjoyable. These learning behaviors a ffect their attitude toward the products and websites, which will determine their purchase decisions ( Lorenzo et al., 2012 ).

All learning implies the integration of two very di fferent processes:

an external interaction process between the learner and his or her so- cial, cultural, or material environment, and an internal psychological process of acquisition and elaboration ( Illeris, 2003). The external in- teraction process is a social dimension such as participation, commu- nication and co-operation. It serves the personal integration in com- munities and society and thereby also builds up the sociality of the learner. The internal psychological process is a process of integrated interplay between two equal psychological functions involved in any learning, namely the function of cognition, dealing with the learning content, and the emotional or psychodynamic function, providing the necessary mental energy of the process ( Illeris, 2003). Thus, the internal psychological process can be divided into two dimensions: the cognitive dimension and the emotional dimension. Many learning theories deal only with one of these processes, which of course does not mean that they are wrong or worthless, as both processes can be studied sepa- rately. However, it does mean that they do not cover the whole field of learning. Actually, the internal psychological process is determined by the external interaction process. In the current study, the external interaction process refers to cus- tomers ’interaction with the SCCs, whilst internal psychological process refers to the attitude formulated after their experience with the social commerce websites. To depict the cognitive dimension and emotional dimension of customers ’learning process, we next introduce cognitive appraisal and a ffective appraisal.

2.4. Cognitive appraisal and aff ective appraisal In social learning theory, the cognitive dimension is the dimension of the learning content, which may be described as knowledge or skills and which builds up the understanding and the ability of the learner.

The endeavor of the learner is to construct meaning and ability to deal with the challenges of practical life and thereby develop an overall personal functionality. The emotional dimension is the dimension en- compassing mental energy, feelings and motivations. Its ultimate function is to secure the mental balance of the learner and thereby it simultaneously develops a personal sensibility ( Illeris, 2003).

These two dimensions are always initiated by impulses from the interaction processes and integrated in the internal process of A. Chen et al. International Journal of Information Management 37 (2017) 627–638 629 acquisition and elaboration. Therefore, all cognitive learning is, so to speak,“obsessed ”by the emotions at stake (e.g., whether the learning is driven by desire, interest, necessity or compulsion). Correspondingly, emotional learning is always in fluenced by the cognition or under- standing (e.g., new information can change the emotional condition).

Many psychologists have been aware of this close connection and it has also recently been thoroughly investigated in neurology. Thus, we in- corporate two variables –cognitive appraisals and a ffective appraisals to describe the cognitive dimension and emotional dimension of the social learning process. According to recent research in social psychology and marketing, cognitive appraisals and a ffective appraisals are also two distinct di- mensions of attitude theory. By interacting with a website, customers will form an attitude toward it ( Lee & Kozar, 2009;Lee, Chen, & Ilie, 2012 ). Attitude plays an important role in customers ’purchase deci- sion-making process ( Kempf, 1999).

Cognitive appraisals refer to the utilitarian aspect of attitude, while a ff ective appraisals are evaluations based on feelings, emotions, and gut reactions that individuals experience in relation to an attractive object ( Lee et al., 2012). As hedonic features of information technologies be- come more and more prevalent, recent IS research has devoted more attention to the dual nature of attitude and suggested that a ffective and cognitive appraisals should be captured distinctively in research models. For example, Teeni (2001)indicated that a website should foster both cognitive and a ffective appraisals, so as to build a more accurate representation of customers ’actual behavior. In particular, researchers in the area of website design have pointed out the im- portance of a ffective website design components ( Lee & Chen, 2011).

In social commerce, on one hand, some customers may want to get more social knowledge from SCCs to aid their purchase decision- making ( Hajli, 2012a ). On the other hand, customers can get social support from SCCs through interacting with other customers. Receiving support from others in a community brings warmth to users ’hearts.

Such good experiences satisfy users in their social interactions with their supporters, help ful fill their social needs, and produce a ffective attachment ( Chen et al., 2013 ). SCCs create vivid interactions and communication with customers. They help customers develop a sense of presence in a virtual environment created by a computer through a telecommunications medium, resulting in greater engagement and af- fective appraisal (e.g., happy, good, relaxed, likable, or satisfactory feelings) ( Chen et al., 2017 ;Liang, Ho, Li, & Turban, 2011 ), which trigger customers ’purchase intention. As a result, we incorporate both cognitive and a ffective appraisals to delineate customers ’decision- making process on social commerce websites.

3. Hypothesis development Based on the theories discussed above, individual observes the model engaging in various social behavior and notes the reinforcements received for these behavior (vicarious reinforcement). If the observer values the reinforcements received by the model, then the observer will attempt to replicate the model ’s behavior and obtain similar re- inforcements ( Chen et al., 2017 ). It therefore involves three types of variables: the person, behavior, and the environment, which mutually in fluence each other ( Bandura & McClelland, 1977 ). In the current study, the SCCs provide the environment for customers ’social learning.

Cognitive appraisal and a ffective appraisal depict the person ’s psy- chology. And purchase intention describes customers ’behavior in social commerce. Thus, we argue that the learning experience with SCCs can stimulate a positive attitude from both the cognitive and a ffective perspectives. The positive attitude further triggers the purchase inten- tion on the social commerce website. The research model is shown in Fig. 1. As we noted previously, cognitive appraisal constitutes the utili- tarian aspect of attitude that refers to its value to an individual ( Lee et al., 2012 ). It is evaluated based on beliefs and knowledge structures ( Lee et al., 2012; Van der Heijden, 2004 ).

Contrasting with sellers, customers confront information shortage when they make purchase decision. They may doubt if the information aiding their decision is trustworthy and they may be suspicious of the seller quality and the product quality ( Pavlou & Dimoka, 2008). This situation is worse in the online context where customers cannot touch the real goods or see the face of the sellers. They can only judge the product and seller with the limited knowledge and information they have. Social commerce provides the environment for customers to ob- serve the behaviors of others. Learning from SCCs gives customers ac- cess to social knowledge and experiences through interacting with others in online forums and communities, browsing ratings and reviews posted, and considering others ’recommendations ( Huang & Benyoucef, 2013 ). Thus, the information learned from SCCs can remedy customers ’ information shortage and mitigate the information asymmetry re- garding the purchase decision ( Chen et al., 2017). Specifically, learning from SCCs will help users become familiar with a website ( Hajli, 2015), foster their trust toward the website and the sellers ( Lu, Zhao, & Wang, 2010 ), and reduce their uncertainty about their decisions ( Pavlou & Dimoka, 2008). Firstly, in online forums and communities, customers can share their shopping experiences and communicate with others. Some social commerce website users may not be clearly aware of what they need. By reading others ’posts and communicating with them, customers learn the regulations on and the shopping process of the website. More importantly, they become more clear on what they really need, reduce their ambiguity in the online environment, and increase their knowledge on the shopping decision ( Hajli, 2012a; Lu et al., 2010 ). Secondly, social commerce users can also bene fit from other customers ’ratings and reviews of the products and the sellers ( Zheng et al., 2013 ). By reading these ratings and reviews, individuals make sure whether the product is indeed what they need, whether the seller is credible, and whether the shopping experience is enjoyable.

Thirdly, social commerce websites have combined the recommendation function and social network applications into social recommendations ( Chang & Hsiao, 2013). Social networks enable customers to share re- commendation information with other members of the networks. Based on social interactions, preference similarities, recommendation trust, and social relations, social recommendations provide the most relevant and attractive information to customers ( Li et al., 2013), and such in- formation is better received because of their personalized nature ( Marsden, 2010 ). Learning from the social recommendations can re- duce the search cost and satisfy customers ’personalized needs ( Chen & Shen, 2015 ).

Overall, the three SCCs provide utilitarian knowledge and help customers with their cognition of the shopping decision. In addition, social interaction through SCCs is also fundamental to the development of cognition ( Lee & Kozar, 2009 ).Illeris (2003) proposed that the ex- ternal interaction process between the learner and her social, cultural, or material environment in fluences the internal psychological process of acquisition and elaboration. Thus, we hypothesize:

H1. Learning from forums and communities is positively associated with a customer ’s cognitive appraisal.

H2. Learning from ratings and reviews is positively associated with a customer ’s cognitive appraisal.

H3. Learning from social recommendations is positively associated with a customer ’s cognitive appraisal.

A ffective appraisal is an evaluation based on emotions, feelings, and reactions ( Lee et al., 2012 ). In social commerce, a ffective appraisal measures hedonic experiences such as feeling happy, good, relaxed, likable, or satisfactory with Web interfaces. In social commerce web- sites, a scene with multiple environmental components provides people with rich information and experiences that keep them interested and occupied ( Lee & Chen, 2011). Customers can acquire emotional support during their interactions with a social commerce website ( Chen et al., A. Chen et al. International Journal of Information Management 37 (2017) 627–638 630 2013; Liang et al., 2011). Receiving support from others in a commu- nity brings warmth to users’hearts, make them feel happy, likable and satisfaction. Such good experiences satisfy users in their social inter- actions with their supporters, help ful fill their social needs, and produce a ff ective attachment ( Chen et al., 2013). In another hand, the vivid interactions and communication with other customers can help custo- mers develop a sense of presence in a virtual environment, including telepresence and social presence ( Shin, 2010). Both telepresence and social presence brings good experience to users and make them feel relaxed, which result in greater engagement and a ffective appraisal ( Kahai & Cooper, 2003; Lee & Chen, 2011). Speci fically, through learning from SCCs, customers can observe others ’experience, com- municate with each other, and exchange emotions with one another, all of which generate social support and social presence for customers. The information and emotion exchange function of SCCs can catch custo- mers ’attention during their interaction with the website, which provide telepresence for customers. Thus, we hypothesize:

H4. Learning from forums and communities is positively associated with a customer ’sa ffective appraisal.

H5. Learning from ratings and reviews is positively associated with a customer ’sa ffective appraisal.

H6. Learning from social recommendations is positively associated with a customer ’sa ffective appraisal.

Finally, we posit that cognitive appraisal and a ffective appraisal have positive impacts on a customer ’s purchase intention on the social commerce website. Previous research has proposed a direct relationship between customers ’cognitive appraisal and a ffective appraisal and behavior intention ( Lee & Chen, 2011; Lee et al., 2012; Van der Heijden, 2004 ) including customers ’purchase intention ( Lee & Kozar, 2009).

These relationships are also supported by the technology acceptance model ( Davis, 1989 ) and the theory of reasoned action ( Fishbein & Ajzen, 1980), both of which argue that attitude is one of the most important predictors of customers ’behavior. Firstly, the higher positive attitude/belief towards the website are, the higher perceived value of the website is. When the perceived value (both in utilitarian and hedonic) of the website is high, customers are more likely to pur- chase on this website other than somewhere else ( Wang, Yeh, & Liao, 2013 ). Secondly, both appraisals towards the website a ffect customers ’ attitude when they evaluate products. They compare their a ffective and utilitarian value separately before purchases ( Lee & Kozar, 2009). If the appraisals are high, customers are more likely to purchase in this website other than in somewhere else. In social commerce, customers evaluate both their utilitarian and hedonic values before purchases ( Lee & Kozar, 2009 ). Through evaluating the social knowledge and the a ff ective experiences they get from the SCCs, customers are more aware of the product quality, seller quality and the satisfactory level of the purchase experience ( Chen et al., 2017). Then customers can make their purchase decision more rationally ( Chen et al., 2013; Lee & Kozar, 2009 ). Thus, we hypothesize:

H7. Cognitive appraisal is positively associated with a customer ’s purchase intention on a social commerce website.

H8. Affective appraisal is positively associated with a customer ’s purchase intention on a social commerce website.

To safeguard against rival explanations, we speci fied various po- pulation characteristics which may in fluence a customer ’s purchase intention as control variables. Previous studies has found that male, senior and highly-educated customers are more inclined to make cau- tious decisions, whilst female, young and lowly-educated customers are more inclined to make impulse purchases ( Luo et al., 2012; Wang & Zhang, 2012). Students who have much leisure time may spend more time on learning from SCCs than non-students, which lead to di fferent behavior patterns. Therefore, to examine the relationship be- tween learning from the SCCs, cognitive and a ffective appraisals, and a customer ’s purchase intention, we control for four factors: gender, age, education, and occupation.

4. Methodology 4.1. Measurements Items measuring the learning behavior in three SCCs were adapted from Chen et al. (2017) , which included three formative items (i.e., the frequency, duration, and material quality). According to social learning theory, if the behavior occurs more frequently, and if the duration is longer, and if the learning materials have higher quality, the learners ’ learning outcomes/e ffects would be better. That means frequency, duration and material quality are elements which would determine individuals ’learning outcomes or learning e ffects ( Cenfetelli & Bassellier, 2009 ). In addition, social learning is funda- mentally learning by observing the actions of other people in daily life ( Lorenzo et al., 2012 ).Chavis (2012) documented that, when a beha- vior occurs, whatever follows it, that is, the consequences of the be- havior, can either increase or decrease the frequency, duration, or in- tensity of the behavior. The quality of online information also Learning from forums and communi Ÿes Learning from ra Ÿngs and reviews Learning from social recommenda Ÿons Cogni Ÿve Appraisal A +ec Ÿve Appraisal Purchase Inten Ÿ on H7 H8 H1 H4 H2 H5 H3 H6 ControlsGender Age Education Occupation External Interaction Process Internal Psychological Process Decision-Making Fig. 1. The Research Model.

A. Chen et al. International Journal of Information Management 37 (2017) 627–638 631 determines the effect of customers learning behavior in SCCs ( Wang & Haggerty, 2011). Thus, it is reasonable to measure the social leaning behavior in SCCs through frequency, duration, and material quality. We asked respondents to rate the frequency of their learning behaviors using 7-point scales from 1 (never) to 7 (very frequently), to rate the duration of their learning behaviors using 7-point scales from 1 (never) to 7 (more than 5 h), and to rate the extent of their agreement with a statement of the information quality using 7-point scales from 1 (strongly disagree) to 7 (strongly agree). All items measuring the other three constructs were adapted from the extant literature to suit this research context. Cognitive appraisal and a ffective appraisal were measured using four and five items, re- spectively, adapted from Lee and Chen (2011)andLee et al. (2012) .

Purchase intention is a mature construct in the online shopping domain.

We adapted the three items from Zhang et al. (2011)to assess custo- mers ’purchase intention in social commerce. Since each item mea- suring these three constructs is a re flection rather than elements of the corresponding construct and previous literature treated these three constructs as re flective constructs ( Cenfetelli & Bassellier, 2009 ), the measurements of these three constructs were manipulated re flectively using a 7-point scale with anchors ranging from “strongly disagree ”(1) to “strongly agree ”(7).

Because the data collection was conducted in China, and in order to ensure that the meanings of the questions were accurately captured by both the Chinese version and the English version, a back-translation process was used. This was followed by a pilot study conducted with 30 respondents to judge the applicability of the instrument items.

Following the pilot test, minor changes were made to the wordings of the items used in the survey. The items used in this study are presented in the Appendix A .

4.2. Survey administration We conducted an online survey with the actual users of a social commerce website in China. A survey hyperlink was placed on the forums of a social commerce site ( https://daren.bbs.taobao.com) op- erated by Taobao in China. This website provides customers with social functions such as sharing their shopping experiences, rating or writing reviews about the sellers and products, recommending the products they like to others. Social commerce can be summarized in terms of two central activities: putting social media tools in e-commerce website and/or putting e-commerce in social media platforms ( Huang & Benyoucef, 2013, 2015; Shadkam & O'Hara, 2013 ). Adding social features to existing e-commerce platforms is a low-hanging fruit of social commerce and studies show a clear revenue bene fit( Chen et al., 2017). In this case, Taobao can be categorized into “putting social media tools in e-commerce website ”. Since 2010, Taobao.com have made many e fforts (e.g., bangpai.taobao.com ,daren.bbs.taobao.com ) in evolving to a social commerce site. To ensure all the respondents have shopping experience on Taobao, participants were asked to recall a recent shopping experience with this social commerce website, and complete the questionnaire based on this experience. They were re- quired to report what they bought on this website. The hyperlink could be shared by customers thus could spread through the social network in Taobao. To attract more participants, each respondent received ¥ 5 as a reward for her participation. To ensure that only one response was submitted per respondent, each participant ’s Internet protocol (IP) ad- dress and demographic information were recorded and carefully ex- amined. In total, 254 actual customers of this social commerce website completed the online survey during a 20-day period in June 2015. A pilot test of 20 subjects suggested that at least five minutes were needed to complete the questionnaire. Hence, 11 questionnaires completed in less than five minutes were abandoned, resulting in 243 valid re- sponses. Through these screenshot, we found that 199 participants answered our questionnaire employing the PC and the other 44 answers were from mobile phones/devices. The proportion of customers who bought search goods is 32.1%, while the proportion of customers who bought experience goods is 67.9%. We note that the age of 82.7% the respondents is in the range of 20– 29, which may be aflaw of our sample. So, we conducted a di fference test between the current study and the large-scale national survey to social commerce users executed by Iresearch. The insigni ficant of Chi-square di fferences indicates that the two populations were similar in distribution, which demonstrates that the age distribution of this sample is not a major concern in this study. The sample demographic information is summarized in Table 1.

5. Data analysis and results Following Anderson and Gerbing (1988) ,wefirst evaluated the measurement model and then tested the structural model. SPSS and SmartPLS are employed to do these analyses.

5.1. Measurement model We first tested the self-developed measurement of learning from three SCCs. According to Petter et al. (2007), if the variance inflation factor (VIF) is smaller than 3.3, little multicollinearity exists among these indicators, which is a characteristic of a formative construct. As shown in Table 2, none of the VIF values was higher than 3.3, sug- gesting that multicollinearity was unlikely to be an issue in this dataset.

These results demonstrated that the three constructs were formative. Assessment of the measurement model includes the evaluation of reliability, convergent validity, and discriminant validity. We assessed the reliability of the re flective constructs with Cronbach ’s alpha and composite reliability (CR). As shown in Table 2, all reflective constructs in our study achieved scores above the recommended value of 0.7 for Cronbach ’s alpha and composite reliability ( Chin, 1998). As for the formative constructs, reliability in the form of very high internal Table 1 Descriptive Statistics of Survey Respondents (N = 243).

Category Items Frequency Percentage (%) Survey of Iresearch (%) N = 2800 Chi-square di fferences Gender Male 91 37.4 Null Null Female 152 62.6 Null Age ≤19 25 10.3 5.6 χ 2(2) = 0.075; p = 0.963 20 –29 201 82.7 75.1 ≥ 30 17 7.0 19.3 Education High school and below 21 8.6 6.9 χ 2(3) = 0.043; p = 0.998 Two-year college 82 33.7 25.1 Undergraduate 132 54.3 58.9 Graduate and above 8 3.3 9.0 Occupation Students 113 46.5 Null Null Non-students 130 53.5 Null A. Chen et al. International Journal of Information Management 37 (2017) 627–638 632 consistency of constructs is actually undesirable (Petter, Straub, & Rai, 2007 ).

We assessed convergent validity using the signi ficance of the item loading or weight and the average variance extracted (AVE). As shown in Table 2, for the three re flective constructs, all item loadings were signi ficant at the 0.001 level, and all AVEs were above the required value of 0.5 ( Chin, 1998). For the three formative constructs, all in- dicators ’weights on their formative constructs were signi ficant. All these results suggested there was adequate convergent validity in our measurement model. We assessed discriminant validity of the three re flective constructs using two approaches. First, as suggested by Fornell and Larcker (1981) , we compared the square root of the AVEs and construct cor- relations. Results indicated that all the square roots of the AVEs (di- agonal entries in Table 3) exceeded the corresponding construct cor- relations (none-diagonal entries), suggesting that the measures of each construct correlated more with their own items than with items mea- suring other constructs. Second, we compared the loadings and cross loadings. As shown in Table 4, the items for each construct that loaded on each distinct factor were higher than the cross loading on other factors. Thus, we concluded that the discriminant validity in this study was adequate.

Common methods variance (CMV) can be a major source of mea- surement error for survey studies, especially when variables are latent and measured using the same survey at a single point of time ( Luo, Li, Zhang, & Shim, 2010). To check the extent of CMV, we first examined the correlation matrix among latent variables. We found that only low to moderate correlations existed among the latent variables, indicating the minimal in fluence of CMV. Furthermore, the results of factor ana- lysis revealed that no single general factor accounted for the majority of the variance. Each factor accounted for less than 30.608% of the total variance, and all 3 re flective factors accounted for 75.218% of the total variance. All these tests con firm that CMV is not a major concern in this study.

Non-response bias is estimated by comparing the early respondents (i.e., the first quartile) with the late respondents (i.e., the fourth quartile). This approach is the most commonly used means for non- response error estimation in IS research (Chen & Shen, 2015). By comparing the first quartile responses to the last quartile (n = 61), no signi ficant di fferences in gender (t = 0.547, p = 0.568), age (t = 1.426, p = 0.159), education (t = 1.013, p = 0.315), occupation (t = 0.704, p = 0.484) and product type (t = −0.727, p = 0.470) were found. On the basis of these findings, it is assumed that non-re- sponse bias is not a serious concern in this study.

5.2. Structural model SmartPLS was used to test the structural model because this com- ponent-based approach not only places minimal demands on the sample size and residual distributions but also can handle both formative and re flective constructs ( Henseler, Hubona, & Ray, 2016 ;Yang, Lu, & Chua, 2013 ).

The assessment of the PLS-model ’s quality is based on its ability to predict the endogenous constructs ( Sarstedt, Hopkins, & Kuppelwieser, 2014 ). In this study, the following criteria facilitate this assessment:

coe fficient of determination (R 2), cross-validated redundancy (Q 2), path coe fficients, the e ffect size (f 2) and the standardized root mean square residual (SRMR). The results of R 2show that the model accounted for 56.5%, 27.4%, and 17.6% of the variances in purchase intention, cog- nitive appraisal, and a ffective appraisal respectively, which indicate an acceptable level of explanatory power. The Q 2is a means for assessing the inner model ’s predictive relevance. A Q 2value larger than zero for a particular endogenous construct indicates the path model ’s predictive relevance for this particular construct ( Hair et al., 2014). In this study, the Q 2values for the three endogenous constructs (i.e., cognitive ap- praisal, a ffective appraisal and purchase intention) are 0.182, 0.115 and 0.403, respectively. The f 2is computed by noting the change in R 2 when a speci fic construct is eliminated from the model. Based on the f 2 value, the e ffect size of the omitted construct for a particular Table 2 Psychometric Properties of Constructs.

Construct Item VIF Weight t-value Loading t-value Learning from Forums and Communities (LFC, formative) LFC1 1.491 0.278 14.349 LFC2 1.488 0.284 19.032 LFC3 1.172 0.356 22.449 Learning from Ratings and Reviews (LRR, formative) LRR1 1.198 0.225 5.488 LRR2 1.153 0.173 4.034 LRR3 1.047 0.457 20.747 Learning from Social Recommendations (LSR, formative) LSR1 1.469 0.279 15.243 LSR2 1.312 0.229 8.439 LSR3 1.182 0.393 18.041 Cognitive Appraisal (CA, reflective) CA1 0.843 25.164 α = 0.877 CA2 0.892 62.695 CR = 0.917 CA3 0.860 32.509 AVE = 0.734 CA4 0.831 37.601 A ffective Appraisal (AA, re flective) AA1 0.905 57.587 α = 0.908 AA2 0.919 70.227 CR = 0.932 AA3 0.765 17.890 AVE = 0.734 AA4 0.872 41.364 AA5 0.813 21.800 Purchase Intention (PI, reflective) PI1 0.851 34.299 α = 0.848 PI2 0.901 58.868 CR = 0.909 PI3 0.880 42.590 AVE = 0.770 Table 3 Square Roots of the AVEs versus Correlations.

Mean S.D. PI CA AA LRR LSR LFC PI 5.337 0.936 0.877 CA 5.299 0.896 0.530 0.857 AA 5.187 0.923 0.596 0.685 0.857 LRR 4.237 0.881 0.465 0.488 0.341 N/A LSR 3.782 1.008 0.350 0.354 0.248 0.505 N/A LFC 4.066 1.133 0.298 0.380 0.374 0.473 0.442 N/A Note : The terms in bold indicate the square root of the AVE for each construct, and the terms in plain text show the correlations. Table 4 Loadings and Cross Loading.

Factor AA CA PI AA1 0.8000.202 0.374 AA2 0.8080.218 0.388 AA3 0.8530.350−0.039 AA4 0.8230.200 0.234 AA5 0.7360.266 0.205 CA1 0.427 0.5800.390 CA2 0.251 0.7880.355 CA3 0.267 0.8210.242 CA4 0.321 0.6620.371 PI1 0.172 0.263 0.809 PI2 0.323 0.281 0.769 PI3 0.202 0.401 0.737 Eigen-values 3.673 2.701 2.652 Variance% 30.608 22.509 22.101 Cumulative 30.608 53.117 75.218 Note: Numbers in boldface are the factor loading of each item. All the factor loadings are signi ficant at 0.001 level A. Chen et al. International Journal of Information Management 37 (2017) 627–638 633 endogenous construct can be determined such that 0.02, 0.15, and 0.35 represent small, medium, and large effects, respectively ( Hair et al., 2014 ). As the results shown in Table 5, most of the f 2values are in the medium range except the four control variables. The results of path coe fficients show that six of the eight hypothesized paths are sig- ni ficant. We used the SRMS criterion to assess the model ’s goodness of fi t. In our case the SRMS is 0.059, less than the 0.08 proposed by Henseler et al. (2016) . All these results indicate an acceptable level of explanatory power and a satisfactory model fit.

Only one of the control variables (i.e., gender) had a signi ficant impact on the dependent variable. This is consistent with results from previous research that e-commerce is oriented toward e fficiency, transaction, and “masculinity, ”while social commerce is oriented to- ward social networking, branding, and “femininity ”(Wang & Zhang, 2012 ). That is, women devote more time on social commerce websites and are more inclined to purchase products on these websites ( Carroll, 2008 ). Although age, education, and occupation may in fluence custo- mers ’behavior online, they have no signi ficant e ffect on purchase in- tention in social commerce context. One plausible reason is that social learning is a generalized behavior which would occur nondistinctively in any age, education, and occupation ( Bandura & McClelland, 1977; Chavis, 2012). As shown in Fig. 2, six of the eight hypotheses were supported.

Speci fically, learning from forums and communities was found to be positively related to cognitive appraisal (b = 0.167, p < 0.05) and a ff ective appraisal (b = 0.268, p < 0.01), thus supporting H1 and H4.

Learning from ratings and reviews was also found to have positive in fluences on cognitive appraisal (b = 0.359, p < 0.001) and a ffective appraisal (b = 0.200, p < 0.05), thus supporting H2 and H5. In ad- dition, both cognitive appraisal (b = 0.606, p < 0.001) and a ffective appraisal (b = 0.171, p < 0.05) were positively associated with pur- chase intention, thus supporting H7 and H8. By comparing the path coe fficients of cognitive appraisal and a ffective appraisal, we found that cognitive appraisal weighted more heavily in terms of driving custo- mers ’purchase intention than a ffective appraisal.

Unexpectedly, the relationships between learning from social re- commendations and cognitive appraisal (b = 0.099, p > 0.05) and a ff ective appraisal (b = 0.030, p > 0.05), however, were not sig- ni ficant. Thus, H3 and H6 were not supported. To explore the reason for the insigni ficant results, we further conducted two regressions, in which cognitive appraisal and a ffective appraisal were set as dependent variable separately and learning from social recommendations was set as independent variable. The regression results show that learning from social recommendations has signi ficant in fluence on cognitive appraisal (b = 0.140, p < 0.05) and a ffective appraisal (b = 0.096, p < 0.05), while the signi ficant relationships become insigni ficant when the other two independent variables are included in the model. This result de- monstrates that learning from social recommendations can alone de- termine customers ’cognitive appraisal and a ffective appraisal to some extent, but the in fluences are much lower than the in fluences of learning from forums and communications and learning from ratings and reviews.

6. Discussions 6.1. Discussions of the empirical results The objective of this research was to explore the di fferent roles of three SCCs on customers ’purchase decision-making process. We sought to achieve this objective by integrating the social learning theory and attitude. The results showed that the model had strong psychometric properties and explained most of the variance of purchase intention.

Our study presents a few interesting findings.

First, both cognitive appraisal and a ffective appraisal in fluence customers ’purchase intention. Successful websites have to incorporate hedonic features to meet users ’need for enjoyment, fun, and arousal.

However, cognitive appraisal plays a more important role in driving customers ’purchase intention than a ffective appraisal. This indicates that customers on social commerce websites are rational. They gen- erally make a purchase decision based on their beliefs and knowledge Table 5 The f 2Values.

Path f 2values P Values LFC →CA 0.0270.374 LFC →AA 0.0620.260 LRR →CA 0.1180.024 LRR →AA 0.0320.252 LSR →CA 0.0090.605 LSR →AA 0.0010.943 CA →PI 0.4380.002 AA →PI 0.0340.296 Gender →PI 0.026 0.266 Age →PI 0.0010.883 Education →PI 0.005 0.592 Occupation →PI 0.000 0.969 Learning from forums and communities Learning from ratings and reviews Learning from social recommendations Cogni Ÿve Appraisal A +ec Ÿve Appraisal Purchase Inten Ÿ on Controls Gender (0.110*) Age (0.022, ns) Education (-0.052, ns) Occupation (-0.011, ns) 0.606*** 0.171* 0.167* 0.268** 0.359*** 0.200* 0.099 0.030 R 2= 0.274 R 2= 0.176 R 2= 0.565 Fig. 2.

Structural model testing results.

Notes : ***p < 0.001; **p < 0.01; *p < 0.05 (two-tailed); ns (non-sig- ni ficant); insigni ficant (p > 0.05) paths are depicted in dashed lines.

A. Chen et al. International Journal of Information Management 37 (2017) 627–638 634 structures (Lee et al., 2012; Van der Heijden, 2004 ) rather than pur- chase products impulsively based on emotions and feelings ( Lee et al., 2012 ).

Second, learning from forums and communities signi ficantly a ffects customers ’attitude in both cognitive and a ffective dimensions.

However, it a ffects a ffective appraisal more than cognitive appraisal.

This is consistent with prior research ( Chen et al., 2013; Liang et al., 2011 ) that found users mainly participate in online social networks (communities) based on their interests, and the contents on social network websites can provide social support for users and satisfy their emotional needs. Third, learning from ratings and reviews signi ficantly in fluences customers ’attitude in both cognitive and a ffective dimensions.

However, its impact on cognitive appraisal is stronger than on a ffective appraisal. This is because the contents in ratings and reviews are mainly related to product quality, seller quality, or shopping experience. These contents are more utilitarian and reduce customers ’uncertainty about their purchase decisions ( Pavlou & Dimoka, 2008).

Fourth, contrary to our expectations, learning from social re- commendations has no signi ficant in fluence on either cognitive ap- praisal or a ffective appraisal in the comprehensive model, while the regression results show that learning from social recommendations alone has signi ficant in fluence on cognitive appraisal and a ffective appraisal. This result demonstrates that learning from social re- commendations can determined customers ’cognitive appraisal and a ff ective appraisal to some extent, but the in fluences are much lower than the in fluences of learning from forums and communications and learning from ratings and reviews. One plausible reason is that the in- formation in forums, communities, ratings and reviews is much richer and more vivid than the information in social recommendations. In social commerce websites, a speci fic social recommendation is usually a picture combining a piece of information which persuades other cus- tomers to buy the products. According to the media richness theory ( Kahai & Cooper, 2003 ), the lack of information richness limits custo- mers ’acquisition of social knowledge and social support, and the lack of information vividness limits customers ’perception about telepresence and social presence ( Chen et al., 2017; Liang et al., 2011 ). On the contrary, the information in forums, communities, ratings and reviews contain the overview of products, introduction of sellers, experience of the purchase and after-sale service and so on. The representation of information can be texts, pictures, audios, videos or any other form. So, customers would pay more attention on forums, communities, ratings and reviews, whilst the social recommendations are uncompetitive re- latively. Thus, the unexpected results are reasonable.

6.2. Limitations and future research Our study had several limitations. First, the current study only considered the three SCCs (i.e., forums and communities, ratings and reviews and social recommendations) which had been investigated by Hajli (2012a) ,Hajli (2015) ,Shanmugam et al. (2016) andChen et al.

(2017) . Although some other SCCs (e.g., social advertisement) may have signi ficant in fluence on customers ’behaviors, we did not consider other social commerce components due to the lack of theoretical sup- ports and the parsimony of research model. The SCCs were not mature constructs as yet. In future research, it will be constructive to make more thorough classi fications, defi nitions and measurements of all the social commerce components, and investigate their in fluence on cus- tomers ’behavior empirically. Second, all constructs in our research model were studied retrospectively using self-response scales, which might have room for recall bias. Nevertheless, we set a number of screening criteria to make sure our data re flected the users ’actual be- haviors and perceptions. It would be useful for future research to measure customers ’purchase decisions and learning behaviors by ob- serving their actual behaviors. Third, this research only focuses on in- vestigating the roles of learning behaviors related to SCCs on customers ’attitude and purchase decisions. Learning behavior related to SCCs is still a black box. Future research can explore the learning behavioral patterns in detail and examine what website features a ffect customers ’ learning process. Fourth, the research subjects in this study were cus- tomers of Taobao, which can only represent one type of social com- merce (i.e., putting social media tools in e-commerce website). Future research can investigate the other type of social commerce sites (i.e., putting e-commerce in social media platforms). In addition, most of the information posted by users in social commerce sites can be classi fied as electronic word of mouth (e-WOM), which is very important to the visibility of individuals and businesses seeking exposure on the In- ternet. E-WOM is defi ned as“any positive or negative statement made by potential, actual, or former customers about a product or company, which is made avail-able to a multitude of people and institutions via the Internet ”(Yoo, Sanders, & Moon, 2013). It will be very interesting to investigate the e ffect the valence of e-WOM from the social learning perspective.

6.3. Theoretical contribution This study makes several contributions to research. First, this study increases our understanding about customers ’decision pattern in the context of social commerce. The proliferation of social commerce has transformed customer behavior. Customers make their purchase deci- sions by relying more on contents posted by other customers. But how customers make their purchase decision based on these contents has not been adequately addressed. On one hand, the research model sheds light on how learning from SCCs in fluences customers ’attitude during shopping, which determines customers ’purchase decision in turn. The results of the relationships examined in this study provide speci fic in- sight about customers ’behavior and mental activity, which provides a useful framework for future studies on social commerce. One the other hand, the results suggest both cognitive and a ffective appraisals drive customers ’purchase decisions. Cognitive appraisal plays a more im- portant role in driving customers ’purchase intention than a ffective appraisal. This indicates that customers on social commerce websites are rational. These relationships provide insights on customers ’beha- viors and mental activities, which provides a useful framework for fu- ture studies on social commerce. Second, this study extends the application scope of social learning theory. Rooted in psychology, social learning theory explains the learning that occurs by observing others ’actions. It o ffers a structured approach to dealing with a variety of behavioral concerns in a number of di fferent disciplines and settings. Through a global literature re- trieval in IS field, we found that social learning theory has been em- ployed to explain the enterprise applications di ffusion within organi- zations ( Lorenzo et al., 2012), software training and skill acquisition ( Yi & Davis, 2003 ) and computer-mediated communication in educa- tion ( Tu, 2000 ). However, it was seldom used to explain customers ’ purchase behaviors in the online context except ( Chen et al., 2017).

Marsden (2010) proposes that the essential social psychology in social commerce is social learning: learning from the knowledge and experi- ence of others we know and/or trust. This study veri fies this argument empirically.

Third, more importantly, this study incorporates three SCCs si- multaneously and gives us a more comprehensive understanding on customers ’purchase decision-making process. Most previous literature examined the role of a single social commerce component on customers ’ purchase behaviors. However, in social commerce, the presence of multiple social features that cannot be isolated from each other requires the examination of them together. This study compares the di fferent roles of these SCCs. The results deepen our understanding on the impact of each social commerce component. A. Chen et al. International Journal of Information Management 37 (2017) 627–638 635 6.4. Practical implicationsThis study also has practical implications. First, to motivate custo- mers ’purchase decision, social commerce websites should pay more attention to features that trigger customers ’cognitions, rather than af- fections. Although many successful websites have incorporated hedonic features (e.g., content variety and mysterious and expressive visual aesthetics) to meet users ’need for enjoyment, fun, and arousal ( Lee & Chen, 2011 ), their main function is to attract customers to visit and stay on the websites ( Wirtz, Piehler, & Ullrich, 2013 ). Once a social commerce website attracts visitors, the utilitarian features would play more important role on facilitating customers ’purchase decision. To achieve this goal, the social commerce websites should provide accu- rate and complete information to make customers ’experience more e ffi cacy, design well-directed navigation to make the customers more easily to get desired information and format rational function layout to make customers more comfortable to use the websites. Second, this study provides implications on how managers should allocate resources and e fforts among di fferent SCCs. Currently, most social commerce websites dedicate the most space to social re- commendations ( Li et al., 2013). Our results point out the importance of online forums and communities, as well as product ratings and re- views. Managers should allocate more resources to constructing forums and communities to foster customers ’aff ective appraisal. For example, Xiaomi has placed many e fforts (i.e., economic and human resources) on developing its online community to provide after-sale services, communicate with customers and make customers communicate with each other ( Chen et al., 2013; Wirtz et al., 2013 ). These tactics can stimulate customers ’aff ective attachment and maintain customers ’ loyalty ( Chen et al., 2017). To stimulate customers ’learning behavior in forums and communities, social commerce sites can conduct various recreational activities to attract more participants and make customers stay longer in the forums and communities, set up a reputation system to encourage active participants, and provide rewards to customers for sharing their experiences. Also, managers should distribute more re- sources to constructing ratings and reviews to enhance customers ’ cognitive appraisal. These can be realized by dedicating more space to customer ratings and reviews on the website and increasing the avail- ability and utility of the information by encouraging customers to post more and high quality reviews. For example, the initial customers ten years ago didn ’t trust the sellers on Taobao. However, the company made great e fforts to construct the rating and review function more reliable. One of the policies is that removing the negative ratings and reviews is forbidden on Taobao. The gradual maturity of ratings and reviews in recent years bene fits both the website and customers a lot.

7. Conclusions In this study, based on the social learning theory, we develop a research model to examine how customers ’learning behavior along three main SCCs a ffects customers ’attitude in both cognitive and af- fective dimensions and how such attitude determines customers ’ pur- chase intention, which increases our understanding about customers ’ decision pattern in the context of social commerce and extend the ap- plication scope of social learning theory. The empirical results suggest that cognitive and a ffective appraisals are the main predictors of purchase intention, with cognitive appraisal having a higher predictive power than a ffective appraisal, which in- dicates that social commerce websites should pay more attention to features that trigger customers ’cognitions, rather than a ffections. In addition, learning from forums and communities and learning from ratings and reviews have signi ficant in fluences on both cognitive and a ff ective appraisals, while learning from forums and communities plays a more important role in formulating a ffective appraisal and learning from ratings and reviews plays a more important role in determining cognitive appraisal. Contrary to our expectation, learning from social recommendations has no signi ficant in fluence on either cognitive or a ff ective appraisal. Thus, managers should allocate more resources to constructing forums and communities and ratings and reviews to en- hance customers ’aff ective appraisal and cognitive appraisal, respec- tively.

Acknowledgements This work was supported by a grant from the National Natural Science Foundation of China –China (No. 71502126, 71332001) and this work was also supported by the independent innovation fund from Tianjin University (No. 2017XRG-0023).

Appendix A. Constructs and Measures Learning from Forums and Communities (Chen et al., 2017 ) 1. How often did you view the information in forums and communities prior to your last shopping experience? (Seven-point scale: 1 = “never ”, 7= “several times a day ”).

2. How much time did you spend on acquiring the information from forums and communities prior to your last shopping experience? (Seven-point scale: 1 = “never ”,7= “more than 5 h ”) 3. The forums and communities provided you with the information you needed for your last shopping experience. (Seven-point scale: 1 = “strongly disagree ”,7= “strongly agree ”) Learning from Ratings and Reviews (Chen et al., 2017) 1. How often did you view the information in ratings and reviews prior to your last shopping experience? (Seven-point scale: 1 = “never ”, 7= “several times a day ”).

2. How much time did you spend on acquiring the information from ratings and reviews prior to your last shopping experience? (Seven-point scale: 1= “never ”,7= “more than 5 h ”) 3. The ratings and reviews provided you with the information you needed for your last shopping experience. (Seven-point scale: 1 = “strongly disagree ”,7= “strongly agree ”) Learning from Social Recommendations (Chen et al., 2017) 1. How often did you view the information in social recommendations prior to your last shopping experience? (Seven-point scale: 1 = “never, ” 7= “several times a day ”).

2. How much time did you spend on viewing the information from social recommendations prior to your last shopping experience? (Seven-point scale: 1 = “never, ”7= “more than 5 h ”) 3. The social recommendations provided you with the information you needed for your last shopping experience. (Seven-point scale: 1 = “strongly disagree ”,7= “strongly agree ”) Cognitive Appraisal (Lee & Chen, 2011; Lee et al., 2012 ) 1. The website was e ffective for achieving the goal of your visit. A. Chen et al. International Journal of Information Management 37 (2017) 627–638 636 2. The website was convenient for attaining the goal of your visit.

3. You felt comfortable using the website to achieve the goal of your visit.

4. The website was helpful for achieving the goal of your visit.(Seven-point scale: 1 = “strongly disagree ”,7= “strongly agree ”) A ff ective Appraisal (Lee & Chen, 2011; Lee et al., 2012) Your overall experience with the website was:

1. Happy 2. Good 3. Relaxed 4. Likable 5. Satisfactory (Seven-point scale: 1 = “strongly disagree ”,7= “strongly agree ”) Purchase Intention (Lee & Kozar, 2009; Zhang, Agarwal, & Lucas, 2011) 1. You will buy this product or service and consider this website as your first choice.

2. You intend to purchase this product or service from the website.

3. You predict you would purchase this product or service from the website. (Seven-point scale: 1 = “strongly disagree ”,7= “strongly agree ”) Note : The actual survey in this study was conducted in Mandarin, whilst this Appendix shows the translated version.

References Anderson, J. C., & Gerbing, D. W. (1988). Structural equation modeling in practice: A review and recommended two-step approach. Psychological Bulletin,103 (3), 411 –423 .

Bai, Y., Yao, Z., & Dou, Y.-F. (2015). E ffect of social commerce factors on user purchase behavior: An empirical investigation from renren.com. International Journal of Information Management ,35 (5), 538 –550 .

Bandura, A., & McClelland, D. C. (1977). Social learning theory. Englewood Cliffs, N.J:

Prentice-Hall .

Bandura, A. (1978). Social learning theory of aggression. Journal of Communication,28 (3), 12 –29 .

Carroll, B. (2008). Social shopping: A new twist on e-commerce. Furniture Today,32(1), 20 –81 .

Cenfetelli, R. T., & Bassellier, G. (2009). Interpretation of formative measurement in in- formation systems research. MIS Quarterly,33 (4), 689 –707 .

Chang, T. S., & Hsiao, W. H. (2013). Factors in fluencing intentions to use social re- commender systems: A social exchange perspective. Cyber Psychology, Behavior & Social Networking, 16(5), 357 –363 .

Chavis, A. M. (2012). Social learning theory and behavioral therapy: Considering human behaviors within the social and cultural context of individuals and families. Journal of Human Behavior in the Social Environment ,22 (1), 54 –64 .

Chen, J., & Shen, X. L. (2015). Consumers ’decisions in social commerce context: An empirical investigation. Decision Support Systems,79 (C), 55 –64 .

Chen, A., Lu, Y., Wang, B., Zhao, L., & Li, M. (2013). What drives content creation be- havior on SNSs? A commitment perspective. Journal of Business Research,66 (12), 2529 –2535 .

Chen, A., Lu, Y., & Sumeet, G. (2017). Enhancing the decision quality through learning from the social commerce components. Journal of Global Information Management , 25 (1), 66 –91 .

Chin, W. W. (1998). The partial least squares approach for structural equation modeling .

Modern methods for business research , . Mahwah, NJ, US: Lawrence Erlbaum Associates Publishers, 295 –336 .

Davis, F. D. (1989). Perceived usefulness, perceived ease of use, and user acceptance of information technology. MIS Quarterly,13 (3), 319 –340 .

Dinner, I. M., Van Heerde, H. J., & Neslin, S. A. (2014). Driving online and o ffline sales:

The cross-channel e ffects of traditional, online display, and paid search advertising.

Journal of Marketing Research ,51 (5), 527 –545 .

Fishbein, M., & Ajzen, I. (1980). Predicting and understanding consumer behavior: Attitude- behavior correspondence .Understanding attitudes and predicting social behavior ,.

Englewood Cli ffs, NJ: Prentice Hall, 148 –172 .

Fornell, C., & Larcker, D. F. (1981). Evaluating structural equation models with un- observable variables and measurement error. Journal of Marketing Research,18 (1), 39 –50 .

Hair, J. F., Jr., Sarstedt, M., Hopkins, L., & Kuppelwieser, G. V. (2014). Partial least squares structural equation modeling (PLS-SEM): An emerging tool in business re- search. European Business Review ,26 (2), 106 –121 .

Hajli, N., Lin, X., Featherman, M., & Wang, Y. (2014). Social word of mouth: How trust develops in the market. International Journal of Market Research ,56 (5), 673 –689 .

Hajli, M. (2012a). An integrated model for e-commerce adoption at the customer level with the impact of social commerce. International Journal of Information Science and Management ,16 (Special-Issue 2012 ECDC), 77 –97 .

Hajli, M. (2012b). Social commerce adoption model. Paper presented at the proceedings of the UK academy of information systems conference .

Hajli, N. (2015). Social commerce constructs and consumer's intention to buy. International Journal of Information Management ,35 (2), 183 –191 .

Henseler, J., Hubona, G., & Ray, P. A. (2016). Using PLS path modeling in new technology research: Updated guidelines. Industrial Management & Data Systems,116 (1), 2 –20 . Huang, Z., & Benyoucef, M. (2013). From e-commerce to social commerce: A close look at design features. Electronic Commerce Research and Applications ,12 (4), 246 –259 .

Huang, Z., & Benyoucef, M. (2015). User preferences of social features on social com- merce websites: An empirical study. Technological Forecasting and Social Change ,95 , 57 –72 .

Illeris, K. (2003). Towards a contemporary and comprehensive theory of learning. International Journal of Lifelong Education ,22 (4), 396 –406 .

Kahai, S. S., & Cooper, R. B. (2003). Exploring the core concepts of media richness theory: The impact of cue multiplicity and feedback immediacy on decision quality. Journal of Management Information Systems ,20 (1), 263 –299 .

Kempf, D. S. (1999). Attitude formation from product trial: Distinct roles of cognition and aff ect for hedonic and functional products. Psychology & Marketing,16 (1), 35 –50 .

Kim, S., & Park, H. (2013). E ffects of various characteristics of social commerce (s-com- merce) on consumers ’trust and trust performance. International Journal of Information Management ,33 (2), 318 –332 .

Kim, H.-W., Gupta, S., & Koh, J. (2011). Investigating the intention to purchase digital items in social networking communities: A customer value perspective.

Information & Management ,48 (6), 228 –234 .

Lecinski, J. (2012). ZMOT-Winning the zero moment of truth . Retrieved fromhttp://www.

zeromomentoftruth.com/ .

Lee, Y., & Chen, A. N. (2011). Usability design and psychological ownership of a virtual world. Journal of Management Information Systems ,28 (3), 269 –308 .

Lee, Y., & Kozar, K. A. (2009). Designing usable online stores: A landscape preference perspective. Information & Management ,46 (1), 31 – 41 .

Lee, Y., Chen, A. N., & Ilie, V. (2012). Can online wait be managed? The e ffect of filler interfaces and presentation modes on perceived waiting time online. MIS Quarterly, 36 (2), 365 –394 .

Li, Y. M., Wu, C. T., & Lai, C. Y. (2013). A social recommender mechanism for e-com- merce: Combining similarity, trust, and relationship. Decision Support Systems,5(3), 740 –752 .

Liang, T. P., & Turban, E. (2011). Introduction to the special issue social commerce: A research framework for social commerce. International Journal of Electronic Commerce , 16 (2), 5 –14 .

Liang, T., Ho, Y., Li, Y., & Turban, E. (2011). What drives social commerce: The role of social support and relationship quality. International Journal of Electronic Commerce , 16 (2), 69 –90 .

Lorenzo, O., Kawalek, P., & Ramdani, B. (2012). Enterprise applications di ffusion within organizations: A social learning perspective. Information & Management,49 (1), 47 –57 .

Lu, Y., Zhao, L., & Wang, B. (2010). From virtual community members to C2C e-com- merce buyers: Trust in virtual communities and its e ffect on consumers ’purchase intention. Electronic Commerce Research and Applications ,9(4), 346 –360 .

Luo, X., Li, H., Zhang, J., & Shim, J. P. (2010). Examining multi-dimensional trust and multi-faceted risk in initial acceptance of emerging technologies: An empirical study of mobile banking services. Decision Support Systems,49 (2), 222 –234 .

Luo, J., Ba, S., & Zhang, H. (2012). The e ffectiveness of online shopping characteristics and well-designed websites on satisfaction. MIS Quarterly,36 (4), 1131 –1144 .

Marsden, P. (2010). Social commerce (english): Monetizing social media . GRIN Verlaghttp:// books.google.com.hk/books/about/Social_Commerce_english.html?id= 8HnQ1F6bSz4C .

O ’Fallon, M. J., & Butter field, K. D. (2012). The in fluence of unethical peer behavior on observers ’unethical behavior: A social cognitive perspective. Journal of Business Ethics ,109 (2), 117 –131 .

Pavlou, P., & Dimoka, A. (2008, December 14 –17). Mitigating product uncertainty in online markets: IT and business solutions and research implications. Paper presented at the Proceedings of the 29th International Conference on Information Systems .

Pavlou, P. A., & Fygenson, M. (2006). Understanding and predicting electronic commerce adoption: An extension of the theory of planned behavior. MIS Quarterly,30 (1), 115 –143 .

A. Chen et al. International Journal of Information Management 37 (2017) 627–638 637 Petter, S., Straub, D., & Rai, A. (2007). Specifying formative constructs in informationsystems research. MIS Quarterly,31 (4), 623 –656 .

Rad, A. A., & Benyoucef, M. (2010). A model for understanding social commerce. Paper presented at the 2010 Conference on Information Systems Applied Research .

Riquelme, I., & Román, S. (2014). The in fluence of consumers' cognitive and psycho- graphic traits on perceived deception: A comparison between online and o ffline re- tailing contexts. Journal of Business Ethics ,119 (3), 405 –422 .

Rondán-Cataluña, F. J., Arenas-Gaitán, J., & Ramírez-Correa, P. (2015). Travel buying behavior in social network site users: To buy online vs. o ffline. Journal of Theoretical & Applied Electronic Commerce Research ,10 (1), 49 –62 .

Shadkam, M., & O'Hara, J. (2013). Social commerce dimensions: The potential leverage for marketers. Journal of Internet Banking & Commerce ,18 (1), 1 –14 .

Shanmugam, M., Sun, S., Amidi, A., Khani, F., & Khani, F. (2016). The applications of social commerce constructs. International Journal of Information Management ,36 (3), 425 –432 .

Shin, D. H. (2010). The dynamic user activities in massive multiplayer online role-playing games. International Journal of Human-computer Interaction ,26 (4), 317 –344 .

Shin, D. H. (2013). User experience in social commerce: In friends we trust. Behaviour & Information Technology ,32 (1), 52 –67 .

Teeni, D. (2001). Review: A cognitive-a ffective model of organizational communication for designing IT. MIS Quarterly,25 (2), 251 –312 .

Tu, C. H. (2000). On-line learning migration: From social learning theory to social pre- sence theory in a CMC environment. Journal of Network and Computer Applications , 23 (1), 27 –37 .

Van der Heijden, H. (2004). User acceptance of hedonic information systems. MIS Quarterly ,28 (4), 695 –704 .

Voorveld, H., Smit, E. G., Neijens, P., Segijn, C., & Bronner, F. (2014). Are online buyers driven by o ffline search? The role of online & o ffline media in the purchase process of di fferent types of products. Paper presented at the American Academy of Advertising Conference Proceedings .

Wang, Y., & Haggerty, N. (2011). Individual virtual competence and its in fluence on work outcomes. Journal of Management Information Systems ,27 (4), 299 –334 .

Wang, C., & Zhang, P. (2012). The evolution of social commerce: The people, manage- ment, technology, and information dimensions. Communications of the Association for Information Systems ,31 (1), 105 –127 .

Wang, Y.-S., Yeh, C.-H., & Liao, Y.-W. (2013). What drives purchase intention in the context of online content services? The moderating role of ethical self-e fficacy for online piracy. International Journal of Information Management ,33 (1), 199 –208 .

Weaver, G. R., Treviño, L. K., & Agle, B. (2005). Somebody I look up to: Ethical role models in organizations. Organizational Dynamics,34 (4), 313 –330 .

Wirtz, B. W., Piehler, R., & Ullrich, S. (2013). Determinants of social media website at- tractiveness. Journal of Electronic Commerce Research ,14 (1), 11 –33 .

Yadav, M. S., de Valck, K., Hennig-Thurau, T., Ho ffman, D. L., & Spann, M. (2013). Social commerce: A contingency framework for assessing marketing potential. Journal of Interactive Marketing ,27 (4), 311 –323 .

Yang, S., Lu, Y., & Chua, P. Y. (2013). Why do consumers adopt online channel? An empirical investigation of two channel extension mechanisms. Decision Support Systems ,54 (2), 858 –869 .

Yi, M. Y., & Davis, F. D. (2003). Developing and validating an observational learning model of computer software training and skill acquisition. Information Systems Research ,14 (2), 146 –169 .

Yoo, C. W., Sanders, G. L., & Moon, J. (2013). Exploring the e ffect of e-WOM participation on e-Loyalty in e-commerce. Decision Support Systems,55 (3), 669 –678 .

Zhang, K. Z. K., & Benyoucef, M. (2016). Consumer behavior in social commerce: A lit- erature review. Decision Support Systems ,86 (C), 95 –108 .

Zhang, T., Agarwal, R., & Lucas, H. C., Jr. (2011). The value of IT-enabled retailer learning: Personalized product recommendations and customer store loyalty in electronic markets.

MIS Quarterly,35 (4), 859 –881 .

Zheng, X., Zhu, S., & Lin, Z. (2013). Capturing the essence of word-of-mouth for social commerce: Assessing the quality of online e-commerce reviews by a semi-supervised approach. Decision Support Systems ,56 (1), 211 –222 .

Zhou, L., Zhang, P., & Zimmermann, H.-D. (2013). Social commerce research: An in- tegrated view. Electronic Commerce Research and Applications ,12 (2), 61 –68 . Aihui Chen is an Assistant Professor of College of Management and Economics in Tianjin University. He re- ceived his PH.D from the Huazhong University of Science and Technology in 2014. His research focuses on social commerce, social network, electronic and mobile business.

His research has been published in Journal of Management Information Systems, Journal of Information Technology, Journal of Business Research and others .

Yaobin Luis a specially appointed professor in information systems and the associate dean of the School of Management at Huazhong University of Science and Technology in China. He received his PH.D from the Huazhong University of Science and Technology in 1997.

His research interests include social commerce, technology implementation, individual adoption, customer trust, elec- tronic commerce, and related topics. He is the author of more than forty publications in various international jour- nals, such as Journal of Management Information Systems, International Journal of Electronic Commerce, Decision Support Systems, Information Systems Journal, Information- & Management, Computers & Education, Electronic Commerce Research and Applications , and others.

Bin Wangis a Professor of College of Business and Entrepreneurship, University of Texas Rio Grande Valley.

Her research focuses on the performance of IT firms, elec- tronic commerce, and IS and economics. Her research has appeared in Journal of Management Information Systems, Information Systems Journal, Electronic Commerce Research and Applications, Electronic Markets, International Journal of Electronic Business, International Journal of Services, Economics and Management, and others .

A. Chen et al. International Journal of Information Management 37 (2017) 627–638 638