The effects of live streaming attributes on consumer trust and shopping intentions for fashion clothing

Abstract Small individual sellers and retailers use live streaming as a direct selling channel to demonstrate and sell their products. This study proposes a framework that examines the influence of live streaming attributes on customer trust and intentions to watch and purchase fashion clothing. Drawing on prior literature, we examine an extensive list of 20 live streaming attributes including product attributes, seller attributes, and other related attributes. The study is performed on 476 Thai consumers with diverse demographics. Results show that product quality and price transparency significantly influence customer trust and intentions to watch and purchase, while seller’s image of being trustworthy and the quality of seller’s Facebook page only show weak relationships. Another finding is that seller pre-announcing their broadcast timing will encourage higher intention to watch. And as expected, the trust in seller positively influences trust in product. These findings suggest opportunities for sellers to focus their attention on important live streaming attributes to develop trust with their customers and increase their customer intentions to watch and purchase. The study concludes with discussion on managerial implications and future work on live streaming commerce for fashion clothing products.


PUBLIC INTEREST STATEMENT
Many small individual sellers and retailers use live streaming as a direct selling channel to demonstrate and sell fashion clothes. This paper examines as many as 20 attributes, such as the seller's presentation skills, the product quality, and pricing, to see how they affect the consumer trust and intentions to watch and purchase. The study is performed on 476 Thai consumers with diverse demographics. Findings suggest that product quality, price transparency, and the pre-announcement of the broadcast timing have significant influence while seller's image of being trustworthy and the quality of seller's Facebook page show weak relationships. These findings also suggest opportunities for sellers to focus their attention on important attributes to develop trust with their customers and increase their customer intentions to watch and purchase. The paper concludes with discussion on managerial implications and future work on live streaming commerce for fashion clothing products.

Introduction
Today, the usage of live streaming as a direct selling channel is growing in popularity. Live streaming is a broadcasting of real-time online videos where a person broadcasting the content is called a streamer. In live streaming selling of goods such as fashion clothing, streamers broadcast content related to the goods being sold and audience usually interact with the streamers and other audience via text chat. Streamers may try on the clothes and describe their properties and audience may interact with the streamers by asking questions, expressing opinions, or making purchase. Many small individual sellers use live streaming feature on Facebook to demonstrate products and conduct sales . To become successful, fashion goods sellers should understand what factors they need to focus on to increase their live stream views and sales.
A number of prior studies have examined consumer behavior in live streaming shopping environments.  has studied perceived shopping values in live streaming and how they enhance trust and in turn lead to customer engagement. Cai et al. (2018) has examined certain live streaming attributes such as seller physical attractiveness and product information, while Hou et al. (2019) has examined factors such as seller interactivity, seller humor, and seller sex appeal. Sun et al. (2019) has examined attributes such as seller's presentation skills, interactivity skills, and shopping guidance. However, the prior research is lacking in its studies of the possible shopping attributes, especially in the context of live streaming shopping for fashion clothes. Only Chandrruangphen et al. (2021) has explored some live streaming attributes in fashion clothes shopping through customer interviews.
Therefore, the study's main objective is to investigate the live streaming attributes that motivate shoppers to watch and shop fashion clothes on live streams. In the following sections, we first review the background of live streaming attributes for fashion clothing shopping, before constructing a conceptual model based on the theories that describe the influences of live streaming attributes on consumer trust and consumer shopping behaviors in live streaming.

Live streaming attributes for fashion clothing shopping
There have been a lot more studies in how traditional store attributes and online shopping website attributes affect consumer shopping intentions than those of live streaming shopping. Though the attributes are not the same, the formers share a lot of common attributes with the latter.
With regard to online stores, Liang and Lai (2002) describes six motivation factors that affect how consumers choose stores (e.g., online order, search function, easy to sign up, home delivery, credit card payment, shopping cart feature), two hygiene factors (security and consistent style), and two media richness factors (e.g., product organization and navigational links). Chen et al. (2010) studies three related areas of website attributes: technology (e.g., security, privacy, and usability), shopping (e.g., convenience, trust, and delivery), and product (e.g., product value and merchandising). Other factors also include product assortment, product quality, price transparency, website convenience and product assortment (Davari et al., 2016;Kautish & Sharma, 2019).  In live streaming shopping,  studies factors that influence consumer trust and shopping engagement. Some of the important shopping attributes being studied in prior literature include seller physical attractiveness, seller interactivity (Cai et al., 2018); streamer interactivity, humor, sex appeal (Hou et al., 2019); seller's product presentation, ability to answer questions, shopping guidance skills (Sun et al., 2019); product information, product interactivity, communication quality, enjoyment, trend setting, and social presence . Chandrruangphen et al. (2021) finds additional factors including seller pacing or the speed at which seller moves from one item to the next, product personal appeal, price transparency, background ambiance, broadcast timing announcement, number of viewers, and seller Facebook page. Based on prior works, we summarize the list of shopping attributes in live streaming that could motivate shoppers to shop as shown in Table 1.

Trust in seller & trust in product
In e-commerce, trust is defined as the beliefs in something or someone based on their characteristics, such as goodness, fairness, honesty, competence, and many others (McKnight & Chervany, 2001). Trust in seller is defined as the customer beliefs in the seller based on seller competency and reliability to serve customer long-term interests (Crosby et al., 1990). Swan et al. (1999) finds that trust in salesperson creates successful sales relationship through positive customer attitudes, intentions, and behaviors. Trust in product is defined as the customer beliefs that the product will meet their expectations ).

Customer engagement
Many studies have examined the topics of customer behaviors in shopping. In both online and offline context, customer intention to purchase has been studied extensively (Aghekyan-Simonian et al., 2012;Davari et al., 2016;Johnson et al., 2015). In live streaming shopping, Cai et al. (2018) and Sun et al. (2019) have studied factors that influence customer intention to purchase.  and Hou et al. (2019) have explored factors that influence customer trust and intentions.

Relationship among live streaming attributes, trust, and customer engagement
Based on prior studies, there are several live streaming attributes that might have positive influence on trust in product, trust in seller, and intention to watch the live streaming. These are as follow: Seller image refers to the customer perception of the seller and the impression of what they expect from the seller. Aghekyan-Simonian et al. (2012) points out that product brand image and online store image reduces risk of online shopping which increases purchase intention. Leeraphong and Sukrat (2018) finds that seller reputation affects customer shopping intentions. Therefore: H1a/b/c. Seller image has a positive influence on trust in product/trust in seller/intention to watch.
Seller interactivity refers to the ability of seller to communicate with shoppers. By enabling shoppers to interact with seller, customer has more trust in the seller and trust in the product, and in turn affect engagement with the seller . Hou et al. (2019) also finds that streamers interacting with viewers affect the viewer intention to continue watching. Therefore: H2a/b/c. Seller interactivity has a positive influence on trust in product/trust in seller/intention to watch.
Seller presentation refers to the ability of seller to present products to shoppers. Sun et al. (2019) shows that product presentation in live streaming increases consumer engagement and shopping intentions. Therefore: Seller shopping guidance is the service given by knowledgeable salesperson to help guide shoppers to find desired products (Darian et al., 2001). Lee and Dubinsky (2017) suggests that online customers prefer to be assisted by salesperson and are likely to buy the recommended products. Sun et al. (2019) shows that personalized product recommendations affect consumer shopping intentions. Therefore: H4a/b/c. Seller shopping guidance has a positive influence on trust in product/trust in seller/ intention to watch.
Seller politeness refers to how much the shopper thinks the seller is a polite person. Shoppers are more inclined to trust and shop with sellers who are likable, friendly, and polite (Bateman & Valentine, 2015;Cai et al., 2018;Nicholson et al., 2001). Therefore: H5a/b/c. Seller politeness has a positive influence on trust in product/trust in seller/intention to watch.
Seller verbal attractiveness refers to how well the seller can talk to keep viewers engaged. Fraser et al. (2019) and Hennig-Thurau (2004) suggest that the ability of streamers to socialize with the audience helps build relationship and influences business success. Therefore: H6a/b/c. Seller verbal attractiveness has a positive influence on trust in product/trust in seller/ intention to watch.
Product assortment refers to availability of products in various qualities, styles, and sizes being sold (Bauer et al., 2012). Product assortment provides value to shoppers in terms of product variety and depth and breadth of selections which has positive effect on trust and shopping intentions (Kautish & Sharma, 2019;Rubio et al., 2017). Therefore: H7a/b/c. Fashion product assortment has a positive influence on trust in product/trust in seller/ intention to watch.
Product quality refers to the superiority or excellence of a product (Zeithaml, 1988). Chinomona et al. (2013) finds that perceived product quality positively influences customer trust and purchase intention. Additionally, stores that offer products of low quality would lose customer trust (Jarvenpaa et al., 2000). Therefore: H8a/b/c. Fashion product quality has a positive influence on trust in product/trust in seller/intention to watch.
Product trendiness refers to the novelty and uniqueness of the products (Workman & Kidd, 2000). Ladhari et al. (2019) finds that young online female shoppers are attracted to trending products. Melewar et al. (2017) suggests that trendiness and innovation are related to brand trust, credibility, and loyalty. Therefore: H9a/b/c. Fashion product trendiness has a positive influence on trust in product/trust in seller/ intention to watch.
Product brand name can be defined as the beliefs or attachments customers have about the brand (Wood, 2000). Web stores with reputable brands are associated with higher levels of brand trust (Ha, 2004). Ladhari et al. (2019) also finds that shoppers see brand value as implying higher trust towards well-known brands. Therefore: H10a/b/c. Product brand name has a positive influence on trust in product/trust in seller/intention to watch.
Product personal appeal is a measure of how clothing items carried by the seller appeal to the unique fashion taste of the shoppers. Ladhari et al. (2019) and Bento et al. (2018) suggest that women who shop fashion clothes follow brands that resonate with their fashion style. Customers trust judgment of salespersons who have unique personal style and fashion taste (McColl et al., 2013). Therefore: H11a/b/c. Product personal appeal has a positive influence on trust in product/trust in seller/ intention to watch.
Pricing transparency is a measure of how pricing information is being communicated clearly. Davari et al. (2016) views that price transparency influences how customers perceive the quality of online stores. Therefore: H12a/b/c. Pricing transparency has a positive influence on trust in product/trust in seller/intention to watch.
The number of live stream viewers refers to how many viewers are watching the live stream. Shoppers feel that high number of viewers signifies a type of social proof that may indicate trustworthiness of seller or product (Chandrruangphen et al., 2021). Wang et al. (2020) shows that the number of viewers affect audience engagement. Therefore: H13a/b/c. The number of viewers has a positive influence on trust in product/trust in seller/ intention to watch.
The seller Facebook page refers to how well the page provides information about the seller and the products. Ruiz-Mafe et al. (2014) finds that users who perceive the FB fanpage of a brand being useful and who have high trust towards the brand will develop higher brand loyalty. Therefore: H14a/b/c. Seller Facebook page has a positive influence on product trust/seller trust/intention to watch.
Seller humor refers to the ability of seller to amuse the audience. Imlawi and Gregg (2014) and Hou et al. (2019) find that humor positively influences social network engagement and increases intention to continue watching. Therefore: H15. Seller humor has a positive influence on intention to watch.
Seller sex appeal refers to the physical attractiveness of the sellers. Cai et al. (2018) and Hou et al. (2019) show that seller physical attractiveness would motivate customers to watch live stream. Therefore: H16. Seller sex appeal has a positive influence on intention to watch.
Seller pacing refers to the appropriate speed in which the seller moves from one item to the next. If seller stays on a certain item for too long, the shoppers will feel bored and may leave the live stream (Chandrruangphen et al., 2021). Therefore: H17. Seller pacing has a positive influence on intention to watch.
Background ambiance refers to how shoppers perceive the environment seen in the background. El Hedhli et al. (2017) and Albayrak et al. (2016) suggest that ambiance in shopping malls affects customer willingness to patronize the mall. Therefore: H18. Background ambiance has a positive influence on intention to watch.
Broadcast timing announcement refers to how appropriately the seller announce the live stream schedule to the viewer ahead of time. It is important for shoppers to know when the seller would broadcast the live stream because they may need to manage their time (Chandrruangphen et al., 2021). Therefore: H19. Broadcast timing announcement has a positive influence on intention to watch.
As for the last remaining live streaming attribute, product pricing in online shopping has been well-studied and plays an important role in how shoppers behave (Grewal et al., 2003). Shoppers tend to shop expensive products that have well-known product brands and with well-known retailers (Forsythe & Shi, 2003). Leeraphong and Sukrat (2018) finds that pricing advantage may influence viewers to make impulse purchases. Therefore: H20a/b/c. Product pricing has a positive influence on trust in product/intention to watch/intention to purchase.
As Huang (2015) shows that product evaluation blogs increase trust in product, it can also be considered that products carried by trusted sellers could be more trusted. Escobar-Rodríguez and Bonsón-Fernández (2017) and Shareef et al. (2019) suggest that trust in online shopping may increase customer purchase intentions. Therefore: H21a/b/c. Trust in seller has a positive influence on trust in product/intention to watch/intention to purchase.
Customers who are satisfied with the product will trust the product and will lead them to purchase the product (Chinomona et al., 2013). Therefore: H22a/b. Trust in product has a positive influence on intention to watch/intention to purchase.
As customers continue to explore more information about the product and receive more information about the product, they could be induced into making purchases (Babin et al., 1994). Therefore: H23. The intention to watch has a positive influence on intention to purchase.

Research model and hypotheses
Figure 1 draws on the above literature and adapted from the trust model in  to present a conceptual framework showing live streaming attributes that influence shopping intentions through trust.

Sampling
Data were collected through an online survey in Thailand. To reach live streaming shoppers, the questionnaire was advertised on Facebook for 10 days. Because we encouraged our respondents to complete the questionnaire, we offered to donate 20 Thai Baht for each completed questionnaire to a charity as a virtuous incentive in this study. The population includes all Thai Facebook users that have used live streaming feature to watch and shop for fashion clothes. Though the accurate number of the population cannot be determined, similar studies such as that of  has shown that the effective sample size using PLS-SEM could be between 150 and 246, where the minimum is based on 10 times the largest construct (Barclay et al., 1995) and the recommended average size is based on 246 (Shah & Goldstein, 2006). We collected the total of 476 Thai respondents. Of this total, 93% (n = 442) had made purchase and 58% (n = 276) were female. Most respondents aged over 36 (n = 185;38.87%) followed by 26-35 (n = 160;33.61%), then by those under 25 (n = 130;27.31%). Most of the respondents were singles (n = 329;69.12%), had a bachelor's degree (n = 268;56.30%), had an average monthly income between 15,001 and 30,000 Thai Baht (n = 196;41.18%), worked as government employees (n = 198;41.60%), and lived in Bangkok (n = 88;18.49%).

Questionnaire and measures
Respondents were required to answer the screening question to ensure they have experience watching or making fashion clothing purchase through Facebook live streaming. The questionnaire was divided into three parts. The first part collected demographic data of the respondent. The second part included the measure of all live streaming attributes. An 8-item measure of product assortment was adapted from Davari et al. (2016) and Kautish and Sharma (2019). A 7-item measure of product quality was adapted from Davari et al. (2016) andEl Hedhli et al. (2017). A 3-item measure of product trendiness and a 3-item measure of product brand name were adapted from El Hedhli et al. (2017). A 4-item measure of product pricing was adapted from Johnson et al. (2015) and El Hedhli et al. (2017). A 3-item measure of product personal appeal was created from Chandrruangphen et al. (2021). A 15-item measure of seller image were adapted from Cai et al. (2018) and Aghekyan-Simonian et al. (2012), which in turn adapted from Vázquez et al. (2002). A 7-item measure of seller interactivity was adapted from Hou et al. (2019). A 4-item measure of seller presentation and 4-item measure of seller shopping guidance were adapted from Sun et al. (2019). A 5-item measure of seller politeness was adapted from Bateman and Valentine (2015) and Cai et al. (2018). A 4-item measure of seller verbal attractiveness was created from Chandrruangphen et al. (2021). A 7-item measure of seller humor was adapted from Hou et al. (2019) and . A 6-item measure of seller sex appeal was adapted from Hou et al. (2019) and Cai et al. (2018). A 2-item measure of seller pacing was created from Chandrruangphen et al. (2021). A 4-item measure of price transparency was adapted from Davari et al. (2016). A 4-item measure of background ambiance was adapted from El Hedhli et al. (2017). A 3-item measure of broadcast timing announcement, a 2-item measure of number of viewers, and a 5-item measure of seller's FB page were created from Chandrruangphen et al. (2021). And lastly, the third part included the measure of trust and consumer intentions. A 3-item measure of product trust, and 4-item measure of seller trust were derived from . A 3-item measure of intention to watch was derived from Hou et al. (2019) and a 3-item measure of intention to purchase was derived from Sun et al. (2019). Among this, the second and third parts of the questionnaire adopted a seven-scale Likert scale, with (1) representing total disagreement and (7) representing total agreement. Since all the respondents were Thais, the questions were developed in English and then translated from English to Thai.

Results
The PLS-SEM analysis was performed using SmartPLS software. The measurement model was used to test the reliability and the validity of the constructs, and the structural model was used to test the hypotheses.

Measurement model
The reliability of the constructs was tested using the individual loadings, composite reliability (CR), Cronbach's alpha, and average variance extracted (AVE) (see, Table 2). To assess the reliability of the individual items, indicator loadings to be kept are at least 0.700. As a result, eleven items were dropped from the analysis (see Appendix). Final set of measurement items is shown in Table 2 along with the values of Cronbach's alpha, and CR to be above 0.8 indicating sufficient internal consistency. The convergent reliability was tested using AVEs for all the factors to be above 0.5 and CR to be higher than AVE, indicating adequate validity. The discriminant validity was tested using the heterotrait-monotrait ratio of correlations (HTMT) to be less than 0.9 and satisfied the Fornell-Larcker criterion indicating that each construct is distinct from the other constructs as it correlates with its own construct more than with other constructs (see , Tables 3 and 4).

Structural model and hypothesis testing
In the results of structural model as shown in Figure 2, a coefficient of determination (R2) is 0.639 for trust in products, 0.490 for trust in seller, 0.609 for intention to watch, and 0.653 for intention to purchase. This indicates that an adequate level of variability in the outcome of the data can be explained by the model. The paths and factors with p > .05 are omitted for simplicity. Table 5 summarizes all the path coefficients and gives the results of the hypotheses.
With regard to the role that trust in sellers and trust in products have on each other and on customer engagement, our results appear to show some similarities but also some contradictions as compared with . As for the similarity, our results show that trust in sellers could lead to customer behavior in terms of intention to watch (β = 0.308; p < .001) and intention to purchase (β = 0.283; p < .001), supporting H21b and H21c. As for the contradictory, while Wongkitrungrueng and Assarut (2020) finds that trust in products does not directly lead to any customer engagement, our results show that trust in products could directly lead to customer behavior in terms of intention to watch (β = 0.2; p < .01), supporting H22a but does not lead to intention to purchase thus not supporting H22b. Also, interestingly, while Wongkitrungrueng and Assarut (2020) has shown that trust in products leads to trust in sellers, our results show the other way around that trust in sellers could lead to trust in products (β = 0.621; p < .001) supporting H21a.
Our results also find that the pre-announcement of broadcast timing (β = 0.222; p < .001) has significant positive influence on consumer intention to watch the live stream which supports H19. However, pricing effect on consumer intention to watch has low p-value but also low coefficient (β = 0.101; p < .044), suggesting weak positive influence on consumer intention to purchase but not significant enough thus not supporting H20c. Pricing also does not have significant influence on consumer intention to watch, thus not supporting H20b. It is also interesting to note that seller pacing effect on consumer intention to watch has low p-value but also low coefficient (β = −0.104; p < .05) thus not supporting H17. And lastly, the intention to watch could lead to intention to purchase (β = 0.538; p < .001) supporting H23.
As for the remaining hypotheses, no other factors have significant influence on intention to watch, thus not supporting H1c, H2c, H3c, H4c, H5c, H6c, H15, H16, H7c, H8c, H9c, H10c, H11c, H12c, and H18. Interestingly, contradicting to the common belief, the number of live streaming viewers have no significant influence on trust and the intention to watch the live stream, thus not supporting H13a, H13b, and H13c. The seller's Facebook page does not have significant influence on the customer intention to watch, thus not supporting H14c.

Indirect and mediating effects
Although some of the factors do not appear to have any direct effects on the customer intention to watch the live stream, but they may have indirect effects (Hayes, 2009). Therefore, we tested indirect effects using bootstrapping procedure with 5,000 samples feature in SmartPLS. Table 6 shows the results. Only the factors that have significant indirect effects on customer intentions are shown.
As Table 6 shows, while product quality, price transparency, and seller's Facebook page do not have direct effect on intention to watch, each of them has indirect effect through trust in seller Also, interestingly, product quality and price transparency do not have direct effect on trust in product, but each has indirect effect through trust in seller (product quality, CI = 0.044 to 0.225; and price transparency, CI = 0.050 to 0.201). Trust in seller fully mediate the effect of each of product quality and price transparency on trust in product.

Discussion
This study examined the live streaming attributes that motivate shoppers to watch and shop fashion clothes on Facebook live streaming. We showed the relationships between live streaming attributes and the influence they have on consumer trust and intentions to watch and purchase. Our findings revealed how live steaming attributes including product quality, price transparency, seller image, seller Facebook page, seller pacing, broadcast timing announcement, and pricing are associated with customer trust and intentions.
The finding that product quality and price transparency have significant positive influence on trust in seller is consistent with prior studies. Halim et al. (2014) and Chinomona et al. (2013) showed that product quality has positive influence on customer trust and intention to purchase. Mittal and Agrawal (2016), Bertini and Gourville (2012), and Kim et al. (2020) showed that price transparency builds customer trust and enhances customer engagement and purchase intentions.
Also, the finding that seller image and seller Facebook page have weak positive influence on trust in seller is consistent with prior studies. Orth and Green (2009) showed that only some aspects of store image lead to trust (e.g., high service quality and frontline employee benevolence) but not the others (e.g., better pricing and product selections). Jung and Kim (2016) found that not all contents on Facebook page enhances brand trust. Specifically, comments made by brand does not impact customer trust, but posts made by other customers affect customer trust.
Our findings revealed that seller's announcement of the broadcast schedule has a direct positive influence on customer intention to watch. This is due to the fact that some people are busy, so    they need to know about the broadcast ahead of time to arrange time to watch. Consistent with the finding of Swan et al. (1999), who showed that salespersons should take their time with the customers explaining each product thoroughly without having the customers feel rushed, our paper found that seller pacing has weak negative relationships with customer intention to watch the live stream suggesting that live streaming sellers should not rush through the products but spend ample time on them.
Additionally, our paper found product pricing to have a weak direct positive influence on customer intention to purchase. This is due to most live streaming sellers offering products at very competitive prices, so customers feel very little incentive to make purchase decision based on pricing alone. And as indicated earlier, the influence on the purchase intention of fashion clothing in live streaming was shown to be dominated by the product quality rather than prices. Also, our findings revealed that customer trust in seller has a direct positive influence on trust in product which is consistent with Swan and Nolan (1985), who showed that salespersons can build trust with their customers through their experience and knowledge of the products, which in turn could positively influence customer attitudes.
Finally, we found that trust in seller and trust in product have a direct positive influence on customer intention to watch and then to purchase. This finding is consistent with Jarvenpaa et al. (2000), who found that customer trust in a store increases the intentions to shop from that store.

Theoretical contribution
This study contributes to the online social commerce research by being among the early studies on live streaming shopping. While focusing on the fashion clothing products, this study is among the first live streaming shopping studies to shed new insights in this product category. We extend recent live streaming shopping studies (Cai et al., 2018;Sun et al., 2019; that involve live streaming values, customer trust, and customer engagement by examining live streaming shopping attributes in affecting customer intentions to watch and purchase.

Managerial implication
This study provides insights that may benefit managers in online social commerce. Live streaming sellers can focus on creating more values in the important attributes to better serve their customers thus enabling higher intentions to watch and purchase. Sellers could carry higher-quality products and ensure that their product pricing is transparent, so that the customers would feel trusted and be more willing to make purchase from the sellers. Also, sellers could plan and preannounce their broadcast timing to give enough time for customers to manage their busy schedule.

Limitations and future research
Since the studies in the area of live streaming shopping is relatively new and still limited in numbers, especially in the areas of fashion clothing products, more research efforts in this area is needed to fully understand its impact on customer behaviors. This study is limited to one platform, Facebook live streaming, one product category, fashion clothing, and one country, Thailand.
In terms of product category, it is possible that the live streaming attributes may have different impacts on customer trust and behaviors in different product categories such as home organizers, fitness accessories, and small kitchen appliances which are more functional than stylish and fun. Moreover, in addition to Facebook live streaming, there are other popular live streaming shopping platforms such as Lazada and Shopee where the nature of users could be different, thus providing different results.
Lastly, the people in Thailand may behave differently in shopping behaviors as compared with shoppers in other countries such as China and the western countries. This means that the different population of the study and other antecedents such as different live streaming attributes could be incorporated into future studies.