How Social Scene Characteristics Affect Customers’ Purchase Intention: The Role of Trust and Privacy Concerns in Live Streaming Commerce

Abstract

(1) Live streaming commerce has refined online consumer engagement by fostering a real-time, socially enriched shopping environment. Despite its growing prominence, the role of social scene characteristics in consumer purchase decisions in live streaming remains insufficiently examined. (2) This study uses the Cognition-Affection-Conation (C-A-C) framework to examine how these social characteristics influence purchase intention through the mediating roles of emotional and cognitive trust, with privacy concerns as a moderator factor. The research employs Structural Equation Modeling (SEM) to test the hypothesis using data from 504 valid responses. (3) The results demonstrate that the characteristics of social scenes proposed in this study enhance consumer trust and positively impact purchase intention. Moreover, privacy concerns weaken the effect of interactivity atmosphere and scene immersion on emotion trust, though they do not weaken the effect of social identity on emotion trust. (4) These findings contribute to the theoretical understanding of live commerce by identifying the psychological mechanisms linking the social service scene to purchasing behavior. They also offer practical implications for platforms and merchants seeking to improve consumer engagement and trust in competitive digital marketplaces. This research highlights the importance of integrating social scenes and privacy management into the strategic design of live streaming commerce services.

1. Introduction

In the current e-commerce environment, TikTok and Douyin, which are two of the most prominent social media platforms, have garnered significant consumer attention by offering highly personalized recommendations and engaging shopping experiences. These platforms enable users to feel a sense of closeness during purchase processes, which notably impacts consumer behavior, increasingly positioning them as the preferred channel for online transactions. Moreover, they have attracted a growing number of merchants who are achieving commercial success, indicating promising growth prospects for businesses operating within these ecosystems [1]. The integration of live streaming into social media has further catalyzed this transformation by enabling real-time interaction between streamers and consumers, directly influencing purchasing decisions. This emerging model blends informative content with social interaction, redefining the traditional boundaries of online commerce [2]. Compared to conventional e-commerce, it introduces novel elements such as the heightened influence of social service environments and ambient atmospheres on consumer decision-making processes [3]. As a result, online sellers now face intensified challenges in capturing and sustaining consumer attention, making the anchoring of purchase intentions a critical strategic objective [4].

In recent years, the literature on live streaming commerce has focused on factors that influence customers’ purchase intentions from different perspectives. The literature indicates that these influences can be categorized into three major domains: physical perspective, psychological elements, and marketing context. For instance, several studies have emphasized the role of physical attributes such as video design [5,6], visual accessibility [7,8], and esthetic presentation [9,10] in driving consumer engagement and purchasing behavior. Currently, other research has underscored the significance of psychological factors, including social presence, customer trust, and entertainment, in shaping consumer intentions. In parallel, contextual marketing variables such as perceived product value, perceived attractiveness, uncertainty, diagnostic value of information, and the features of live e-service environments have also been identified as influential [11,12]. A notable contribution by Qaisar et al. [13] explored the e-service environment in live streaming commerce, demonstrating that timely interaction enhances consumer trust, which in turn fosters purchasing intentions. Although existing research has addressed various antecedents of consumer behavior in live streaming commerce, there remains a critical gap in understanding the role of social scene characteristics that extend beyond real-time communication and interactivity in shaping consumer decisions.

The emergence of mobile video live streaming has disrupted traditional service and marketing paradigms by simulating real-life social scenarios in digital environments. In this context, sellers craft immersive e-service atmospheres that resonate with consumers, creating environments that not only facilitate product presentation but also fulfill consumers’ social and emotional needs. According to Maslow’s demand motivation theory, when consumers’ social needs are met, they exhibit high levels of participation, ultimately stimulating them to make purchases [14]. Unlike traditional e-commerce, consumers can satisfy their social needs through real-time interactions with other casters in live streaming commerce [15]. However, service scenario factors that can meet social needs go beyond just communication or real-time interactions. Drawing on practices from both live streaming commerce and traditional service scenario theory [16], this study identifies three core components of social scenes: interactivity, scene immersion, and social identity. When consumers feel engaged, acknowledged, and responded to within these environments, they become more emotionally invested, leading to heightened enjoyment and a stronger inclination to purchase [17]. Moreover, mutual recognition among participants fosters a renewed sense of self-awareness and identity, enabling users to express emotions and alleviate psychological pressures, thereby satisfying deeper social needs [18].

According to the Cognition-Affect-Conation (C-A-C) framework, when consumers recognize the social characteristics of service scenarios, this can increase their emotional and cognitive trust, thereby promoting purchases. Additionally, the effect of social scene characteristics has an impact on trust, which is also influenced by another psychological factor, namely privacy concerns. Users’ concerns about privacy breaches in online social scenes are also factors that need to be considered in the seller’s creation of live e-service scenes [19]. The existing literature fails to address how their social components operate through trust mechanisms or how they may be affected by emerging consumer concerns, such as privacy. Therefore, this study addresses this gap by developing a comprehensive model that explains how social scene characteristics influence consumer purchase intentions, and how this relationship is mediated by customer trust and moderated by privacy concerns. Finally, this study shows three important contributions. First, it extends the applicability of the Cognition-Affection-Conation framework to live streaming commerce, offering a nuanced lens through which to examine user behavior on platforms like TikTok and Douyin. Second, we extend previous studies and reveal the psychological mechanisms linking the social scene through which the social characteristics of the live streaming scene, such as interactive atmosphere, scene immersion, and social identity, influence the shopping intention via two distinct types of consumer trust, cognitive and emotional. Lastly, this study contributes to the literature about the privacy concerns in live streaming by presenting robust empirical evidence that privacy concerns moderate the effect of social scene characteristics on consumer trust, offering new theoretical and practical insights into how platforms and merchants can optimize user engagement when they build a social scene and offer a better e-scene service in an increasingly privacy-conscious digital landscape.

2. Theoretical Basis and Hypothesis Development

2.1. Cognition-Affect-Conation Framework

Cognition refers to the perceptions, knowledge, and information that are formed through multiple experiences [20]. Affect encompasses emotional responses or sentiments, while conation pertains to a person’s motivation to act or engage in specific behaviors [21]. The Cognition-Affect-Conation framework offers a robust theoretical foundation for understanding the emotional and intellectual processes underlying consumer decision-making. It has been effectively employed to examine the relationships among perceived value (cognition), emotional satisfaction (affect), and motivational commitment (conation), particularly in mobile and digital environments. Furthermore, this tripartite model, often referred to as the “Trilogy of the Mind”, captures the essence of human consciousness across three core dimensions: cognition, which enables individuals to acquire and process information; affect, which reflects emotional responses; and conation, which translates cognition into goal-directed actions [22]. Previous studies have used the C-A-C framework to analyze user behaviors in the context of social media, verifying the explanatory power of the framework to the sequential causal relationships among cognition, affect, and conation [23]. This study adopts the cognition-affect-conation framework to explain the relationship between variables in the context of live streaming commerce on TikTok or Douyin. Lee et al. [24] emphasize that repeated exposure to an object strengthens an individual’s confidence, as it offers more opportunities for information processing, leading to reinforced attitudes and heightened perception of the associated product, brand, or service. Also, when customers make purchasing decisions, their trust and attitudes are affected by information, knowledge, and perception about the product and service [25,26,27].

Social identity is inherently connected to cognition, as it reflects individuals’ identification with group culture and viewpoints, and involves a re-evaluation of their self-identity and role within highly interactive communities [28]. Good social cognition can meet customers’ social needs and foster greater emotional satisfaction and customer trust, generating a deeper bond with the experience in the live e-service atmosphere, therefore reinforcing the purchase intention. Another variable, the immersion scene, according to Li et al. [29] can be interpreted into the cognitive approach as it has the ability to trigger an effective response by increasing the perception of social and emotional presence in the audience since immersive experiences generate a stronger connection with the content, which in turn influences the purchase intention through emotional satisfaction and the perception of closeness. Furthermore, from the cognitive approach, the cognition of scene immersion can also stimulate information processing, such as visual quality and scene features, as these help consumers better understand the product or service scene, directly affecting their cognitive evaluations of the value and trust. The cognition of the interactive atmosphere fosters greater emotional satisfaction and cognitive trust, which reinforces the purchase intention by providing an experience for direct interaction and fitting customers’ social needs well [30].

Regarding the affective factors, there are two different types of trust: cognitive trust and emotional trust. Cognitive trust in the study is trust based on the rational perception of sociality in the service scenarios provided by the seller. Emotional trust implies an affective connection with the sellers [31]. Emotional trust based on familiarity, acceptance, and mutual understanding can motivate stronger purchasing connotations [32]. The conation factor, which is related to the intention or motivation to act in this study, is the purchase intention. This variable reflects the purchase intention, which is a direct manifestation of the consumer’s willingness to act.

Finally, privacy concerns are introduced as a moderating variable that has not yet been sufficiently examined in live streaming commerce. Privacy concerns relate to consumers’ apprehensions about how their personal information is collected, utilized, and safeguarded in digital interactions. Including this variable enhances the explanatory capacity of the model by incorporating contemporary concerns that influence trust and purchase behavior in socially immersive and data-sensitive environments. Overall, the C-A-C framework offers a multidimensional approach to understanding trust formation and purchase intention in live streaming commerce, capturing the interplay among cognitive, emotional, and motivational components within increasingly complex and interactive digital ecosystems.

2.2. Research Assumptions

2.2.1. Social Scene Characteristics on Consumers’ Purchase Intention

According to Csoban-Mirka et al. [33], the fact that there is an interactive atmosphere during the live broadcast can easily be a critical aspect that includes the physical elements or characteristics of the interior of a store such as lighting, aroma, music and the arrangement of the merchandise. Based on this, they propose that the atmospheric characteristics of online stores should be integrated with technological features such as system quality and interface design, as well as personalized functions like fast access and automatic error correction. These elements should work together to enhance user satisfaction by fostering enjoyment, cognitive engagement, user empowerment, credibility, organization, and improved visual presentation of information. Therefore, the interactive atmosphere is considered a key factor that influences the purchase intention of consumers in e-commerce environments, especially in formats such as live streaming, since it is possible to interact in real time, which generates an increase in the trust and emotional commitment of customers. This, in turn, drives purchase intention as it allows users to interact with hosts and other users beyond the limitations of time and space, strengthening consumers’ product knowledge and purchase intention through mutual online visits [34].

In addition, Li and Hua [35], mentioned that interactivity in live streaming can produce a strong social presence and improve the perception of interaction with consumers since live commerce is based on interactions between consumers and sellers to obtain information about products or brands, which can generate strong purchase interest. Finally, Zhang et al. [36] presented another important feature: perceived social presence is strengthened in an interactive environment, generating an immersive experience for consumers and facilitating purchase decision-making. It should be noted that this immersion not only reduces psychological barriers but also mitigates perceived risk, significantly increasing the probability of purchase intention.

The concept of immersion refers to an individual’s active engagement within a simulated environment that momentarily disconnects them from the real world. Thus, immersive technologies blur digital and physical barriers at the same time since they provide immersive sensory experiences, whether visual (sight), auditory (hearing), olfactory (smell), or haptic (touch). Therefore, brands or ventures that are dedicated to electronic commerce use immersive technologies to create attractive user experiences, which can undoubtedly influence consumers’ purchasing decisions. For their part, they show that users mainly trust what they see with their eyes since the better the immersion and entertainment of the platform’s interface, the better the user experience, as is the case with the live streaming environment as users can immerse themselves by receiving instant messages sent by the presenters, participate in social interactions, and even make the purchase during this immersion [37,38].

Zhou and Wang [39] mentioned the importance of social identity as a key feature of live streaming. They point out that social identity theory refers to how individuals define themselves based on two social levels: social category and group membership. This theory is crucial to exploring the connection between individual identity and social groups, and is therefore considered a description of individual characteristics derived from a person’s intention to identify with one or more social groups. From an economic standpoint, Deng and Wu et al. [1,40] suggest that social identity can also reflect how consumers identify their preferences and tastes about specific products, brands, or market behaviors. This identification process influences purchasing motives, consumption habits, and brand selection, which, in turn, affect the economic structure and growth pattern of society. Not to mention that consumer social identity involves more factors, including the level of attractiveness generated by online celebrities, which has a positive and significant effect on consumers’ effective and hedonic purchase value. Furthermore, as a streamer’s popularity increases, viewers’ experience of live broadcasts improves, which generates positive emotional attitudes. This, in turn, drives consumers’ desire to continue participating in the broadcasts and promotes their purchase intentions. Thus, I propose the following hypothesis:

H1a. 

Interactive atmosphere positively impacts consumers’ purchase intention.

H1b. 

Scene immersion positively impacts consumers’ purchase intention.

H1c. 

Social identity positively impacts consumers’ purchase intention.

2.2.2. Customer Trust and Purchase Intention

Several studies have highlighted the positive impact of emotional trust on purchasing decisions; for instance, Zhang et al. [15] found that emotional trust, developed through friendly, consistent, and transparent interactions during live streaming, significantly improves customers’ willingness to purchase. Furthermore, emotional trust increases when consumers feel that streamers are honest and approachable, fostering a sense of security and personal connection, ultimately leading to higher purchase intentions. However, emotional trust is built through streamers’ authentic product presentation, personal rapport, and real-time responses to audience questions. This emotional connection often makes up for the lack of physical contact in online shopping, making consumers feel more confident about their purchasing decisions in a live streaming environment [6].

According to Alnaim [41], consumer trust is favorably correlated with online vendor trust. In the current digital security and safety regime, the global menace of social identity theft has escalated. When online shopping, the customer’s top priority is to remain safe. Also, they want their privacy to be respected and left alone. Adequate privacy concerns include several distinct situations: data protection, access control, application security, and network security. If firms in online selling implement multi-level security, it will increase client confidence and positively influence their intent to purchase online. Meanwhile, Tran et al. [42] noted emotional trust as the willingness to accept the vulnerability (risks) of Internet commerce websites after gaining knowledge about these websites in the context of e-commerce. This gives way to emotional trust, which is described as the trust a customer has in an online store based on instincts, intuitions, or feelings triggered by the degree of attention and concern of the online retailer, which leads to purchase intent. Thus, trust not only influences consumers’ purchase intention but also strongly motivates their attitudes, online transactions, and beliefs, limiting uncertainty and increasing consumer knowledge control of uncertainty in online transactions. Thus, I propose the following hypothesis:

H2a. 

Emotional trust has positive effects on customer purchase intention.

Research suggests that cognitive trust positively influences purchase intention by enhancing the perceived trustworthiness of information provided during live streams. According to Liu et al. [43], cognitive trust is generated when viewers perceive the streamer or platform to be knowledgeable and trustworthy, leading to increased trust in the product being sold. This trust is based on rational assessments of the streamer’s ability to deliver on their promises, the accuracy of the product information, and the platform’s reputation for offering genuine products. Similarly, Wongkitrungrueng and Assarut [44] found that cognitive trust in both the streamer’s expertise and the platform’s security features significantly impact consumers’ purchase decisions. Their study showed that when consumers trust the factual content shared during the live stream, they are more likely to feel confident about the transaction, which ultimately increases their purchase intention. Finally, another study by Wu et al. [40] emphasized that cognitive trust increases when transmitters provide clear, consistent, and accurate information about the product; this transparency reduces consumer uncertainty and fosters stronger purchase intention. On the other hand, cognitive confidence is more driven by knowledge and competence and arises from the accumulated experience of users after using any system that allows them to make predictions with a certain level of confidence [45,46]. Thus, I propose the following hypothesis:

H2b. 

Cognitive trust has positive effects on customer purchase intention.

2.2.3. Scene Characteristics, Customer Trust, and Purchase Intention

In terms of scene characteristics, customer trust, and purchase intention, the interaction quality refers to the perceptions consumers have during their contact with service personnel. Key factors in measuring this quality include the attitudes, behaviors, and skills of the staff. Additionally, presenters can respond to consumer questions by analyzing information in real-time, thus meeting individual consumer needs. On the other hand, the higher levels of interactivity during live broadcasts lead to greater emotional trust among viewers. Real-time communication and participation opportunities allow consumers to form more personal connections with streamers. Features such as real-time chat, emojis, and virtual gifts foster emotional bonds and create a shared experience atmosphere, contributing to the development of emotional trust. Additionally, the interactive elements in live streams enhance viewers’ sense of social presence, further strengthening emotional connections and trust [2,47].

Trust is the consumer’s perception of a seller’s or company’s reliability and credibility. A valuable tool for measuring this perception is customer experience measurement, which provides an accurate reflection of the content and scope of consumer perceptions, often referred to as cognitive trust. Building trust through positive customer experiences encourages consumers to interact with streamers, accept the authenticity of their offerings, and perceive the value of live streaming e-commerce. In this context, trust acts as a catalyst for consumer purchase intention in a dynamic and rapidly evolving market. Further, the higher levels of interactivity during live streams lead to increased cognitive trust among viewers. Real-time communication and engagement enable consumers to interact directly with streamers, ask questions, and receive immediate answers, which fosters trust in the streamer’s knowledge and competence. Additionally, interactive features like polls, quizzes, and chat functions create a more engaging environment that supports the development of trust [48]. Meanwhile, on Douyin, the Chinese version of TikTok, certain features enhance the consumer experience in ways not found on Western TikTok. The “Yellow Basket” function enables retailers to include product links in every video they upload, according to Putri et al. [49]. Customers can use this functionality to save, like, and comment on the video in addition to swiftly accessing the product purchasing page. By assisting customers in creating mental maps of sellers’ actions, these interactions strengthen trust even more.

While it is true, as explained above, that the interactive atmosphere of a platform such as TikTok can influence consumers’ purchase intention, Li et al. [50] noted that interactive behavior in the information transmission process reduces psychological distance, improving consumers’ sense of presence and enriching their online shopping experience. According to Zhou and Tong [51] in their study on the factors that influence consumers’ purchase intention during live streaming, focusing on the mediating effect of emotion, they present that the characteristics of the live broadcast scene are a very important point because they could stimulate emotional trust, thus stimulating consumers’ purchase intention. While Han et al. [34] say that consumers’ trust was more due to the interactivity with the host and the atmosphere of the live broadcast, it is clear that interactivity has a significant influence on perceived trust, and perceived trust can trigger consumers’ purchase intention. However, it is worth noting that Li et al. [52], explain that when a user enters a live room for the first time, they can determine the popularity of the live room and the interest in its content based on the number of people who have joined the online room and how people react, which is why the interactivity of the atmosphere is important as this helps to identify that there is a lot of communication between the streamer and the users, making them more interested in staying in the room, which will influence the final purchase intention. Thus, I propose the following hypothesis:

H3a. 

Interactive atmosphere during the live streaming process positively impacts emotional trust, and emotional trust plays a mediating role between the interactive atmosphere and purchase intention.

H3b. 

Interactive atmosphere during the live streaming process positively impacts cognitive trust, and cognitive trust plays a mediating role between the interactive atmosphere and purchase intention.

Regarding scene immersion, research by Liu and Zhang [53] revealed that more immersive live streaming experiences, with high-quality video and audio, generated positive effects towards streamers, as the immersive nature of live streaming allows viewers to feel more present in the streamer’s environment, creating a sense of intimacy and shared space. This greater sense of presence and ability to observe the streamer’s authentic reactions and emotions can contribute to higher levels of emotional trust. Likewise, Cho [54] found that immersive elements such as 360-degree views and augmented reality features in live streams can improve viewers’ emotional engagement and trust as they generate a greater sense of immersion for the consumer. Furthermore, Zhang et al. [15] mentioned that, if the consumer is accustomed to traditional online shopping, participating in live streaming could generate a sense of immersion. Therefore, it is worth noting that one of the most widely used models for this type of online commerce studies is the Stimulus(S)-Organism(O)-Response(R) model [55] because it frequently has to do with how people make decisions. Because live streaming is a medium that facilitates two-way information exchange between buyers and sellers, it can generally produce a stronger sense of social presence than one-way media. As a result, interaction is a process of emotional accumulation, and the high social presence created by emotional representation in social situations will increase the consumer’s level of trust. At the same time, when the external environment stimulates the consumer’s psychological consciousness, this will change and lead to purchase intention.

Thus, I propose the following hypothesis:

H4a. 

Scene immersion positively impacts emotional trust, and emotional trust plays a mediating role between scene immersion and purchase intention.

H4b. 

Scene immersion positively impacts cognitive trust, and cognitive trust plays a mediating role between scene immersion and purchase intention.

Another fundamental characteristic of live streaming commerce is the role of social identity. When viewers perceive a strong alignment between their social identity and that of the streamer, they tend to develop higher levels of emotional trust. This is because they can share interests, values, or demographic characteristics between viewers and streamers, which creates a sense of belonging and emotional connection. This alignment is often rooted in shared interests, values, or demographic attributes, which foster a sense of belonging and emotional connection within the streaming environment. The congruence of social identity enhances the perceived emotional relatability and trustworthiness of the streamer, particularly when the streamer effectively communicates their authentic personality and values. In such cases, emotional trust is more readily cultivated among viewers who resonate with these traits. However, social identity refers to the emotional and value-based meaning attributed to an individual by the specific social group to which they belong; this emotional and value-based meaning can influence consumer purchasing behavior in specific contexts, as it significantly affects how consumers evaluate products, thereby increasing their purchase intentions for items linked to social identity [39,47]. Additionally, Zheng et al. [56] highlight that emotional trust is formed between the trusting and trusted parties based on emotional factors; the trusted party shows concern for the well-being, goals, and intentions of the trusting party, which is facilitated by good communication. So, emotional trust, in turn, influences consumers’ purchase intentions, as they are more likely to choose products recommended by trusted opinion leaders, with their emotional trust impacting their purchasing decisions. Thus, I propose the following hypothesis:

H5a. 

Social identity positively impacts emotional trust, and emotional trust plays a mediating role between social identity and purchase intention.

H5b. 

Social identity positively impacts cognitive trust, and cognitive trust plays a mediating role between social identity and purchase intention.

2.2.4. Privacy Concerns, Scene Characteristics, and Customer Trust

Privacy concerns have been identified as a critical moderating factor that can attenuate the positive effects of interactive and immersive features on customer trust in live streaming commerce. As noted by Peng et al. [31], when privacy risks are perceived as high, the interactive atmosphere that typically fosters engagement and emotional trust is less effective. In such cases, even real-time responses from streamers may fail to build trust if customers fear that their data could be misused. In a similar vein, Lutz and Newlands [57] argue that platforms with a high degree of interactivity and user data collection exacerbate privacy concerns, leading to distrust in the platform. When users feel that their data are at risk, they may engage less with interactive features, which decreases trust and ultimately affects their purchasing decisions.

On the other hand, according to Zhang et al. [15], privacy concerns also negatively affect scene immersion and its influence on trust, showing that, when viewers are immersed in a live stream but also concerned about how their personal information may be used, their emotional engagement and trust levels decrease. While immersion can enhance trust, increased privacy concerns diminish the emotional connection that immersion seeks to build, weakening the trust generated through immersive experiences in live streams. Moreover, social identity in live streaming is often built through a sense of community and shared interests, which can increase trust in both the streamer and the platform. However, if users feel that their participation in these social groups exposes them to privacy risks, they may withdraw from the community, which diminishes their sense of social identity and consequently reduces trust. Privacy concerns often prevent users from fully participating in communities formed around live streaming, as they fear that their data may be compromised. This erodes the trust that is typically built through social identity, as users are less willing to fully participate in group dynamics and trust-building activities [31]. Thus, I propose the following hypothesis:

H6a. 

Privacy concerns negatively moderate the relationship between interactive atmosphere and emotional trust.

H6b. 

Privacy concerns negatively moderate the relationship between scene immersion and emotional trust.

H6c. 

Privacy concerns negatively moderate the relationship between social identity and emotional trust.

H6d. 

Privacy concerns negatively moderate the relationship between interactive atmosphere and cognitive trust.

H6e. 

Privacy concerns negatively moderate the relationship between scene immersion and cognitive trust.

H6f. 

Privacy concerns negatively moderate the relationship between social identity and cognitive trust.

Based on the above discussion, the proposed research model is illustrated in Figure 1. The C-A-C model is well suited to explain how social scene features influence consumer purchase intention. This research is well aligned with the C-A-C model.

3. Methodology

3.1. Data Collection and Sample Characteristics

A total of 800 questionnaires were distributed on-site in person and online (e.g., WJX.CN). The respondents have experience in live streaming on social platforms such as TikTok (for international users) and Douyin (for Chinese users). After reviewing the data, questionnaires that did not have logical answers and were completed randomly were deemed invalid and eliminated; finally, 504 valid samples were obtained, whose features matched the characteristics of the customers participating in the live streaming commerce (see

Table 1. Sample demographics.

Characteristics	Item	Frequency	%
Main residence in the past six months	America (South, North, or Central)	23	4.6%
	Europe	2	0.4%
	Asia	479	95.0%
Gender	Male	209	41.5%
	Female	295	58.5%
Age	Under 18 years of age	5	1.0%
	18–44 years old	481	95.4%
	45–59 years old	18	3.6%
Education level	Below high school.	1	0.2%
	High school, technical secondary school.	7	1.4%
	College, university.	442	87.7%
	Graduate and above	54	10.7%
Employment status	Employed	62	12.3%
	Student or pre-school child	423	83.9%
	Retired	19	3.8%

).

Among the valid respondents, a significant majority of respondents (95.0%) resided in Asia, specifically in China, during the past six months, indicating a more limited but still relevant engagement with live streaming shopping in these regions. The sample was slightly skewed toward female (59.5%) respondents compared to male (40.5%) respondents. In terms of age, the majority of respondents (95.4%) were between 18 and 44 years old, which aligns with the demographic characteristics of typical consumers in live streaming commerce. This age group, although commonly associated with youth, also includes a broad range of adults, suggesting that live streaming shopping appeals to a wider audience beyond just young people. Most respondents (87.7%) had attained a college or university education. The characteristics of the sample can represent the overall characteristics of live streaming customers, and since most of the samples are between 18 and 44 years old, they are generally consistent with the actual age distribution of live streaming customers.

3.2. Development of the Instrument and Survey

To verify the hypotheses presented above about the factors influencing consumers’ purchase intention, a questionnaire was developed. The questionnaire mainly includes two parts: the first part consists of the demographic characteristics of the survey participant, such as primary residence in the last six months, gender, age, education level, and employment status. The second part consists of the measurements of variables, including privacy concerns (PC), cognitive trust (CT), emotional trust (ET), interactive atmosphere (IA), scene immersion (SIM), social identity (SID, and purchase intention (PI). A 5-point Likert scale, with 1 denoting strongly disagree and 5 denoting strongly agree, was used to measure each characteristic. All measurement items are selected based on the previous research literature and adapted appropriately (see

Table 2. Summary of measurement items.

Constructs	Items
Privacy concerns	PC1	I am concerned about how TikTok uses my personal information.	[58,59]
	PC2	I am usually worried when social media platforms ask me to provide personal information.
	PC3	When shopping, I am concerned about the security and threats to my online privacy.
Cognitive trust	CT1	The product information in live streams is reliable.	[40,60]
	CT2	Given my interactions with sellers on the TikTok platform, I don’t mind following their recommendations.
	CT3	I can trust that sellers on TikTok conduct thorough product analysis before selling to me.
Emotional trust	ET1	I feel safe relying on sellers on TikTok for shopping.	[59]
	ET2	I feel satisfied relying on sellers on TikTok for shopping.
	ET3	If I tell the sellers about my problems, I am confident they will give me some advice.
Interactive atmosphere	IA1	Streamers and viewers actively respond to other viewers’ questions.	[8,61,62]
	IA2	Streamers and viewers actively communicate with each other.
	IA3	The live stream interface is interactive, visually appealing, and provides a good experience.
Scene immersion	SIm1	I am immersed in the live streaming environment and forget about time.	[63,64]
	SIm2	I am very focused (immersed) when watching live streams.
	SIm3	I am fully engaged in live shopping.
Social Identity	SId1	In live streams, I can freely express my true emotions and thoughts without worrying about traditional social roles.	[65]
	SId2	Interacting with hosts and other viewers in live streams gives me a sense of community and social identity, reshaping my image and social role.
	SId3	I can freely express my emotions and views in live streams without worrying about external judgment.
Purchase intention	PI1	I intend to purchase products or services on this platform.	[60]
	PI2	I will consider buying products on this medium in the future.
	PI3	I am willing to watch live streams on this media to find products we like.

). Six academics took part in discussing the appropriateness and modifying the items. Then, the final measurement items were determined after a pilot study with 60 online participants.

4. Results

4.1. Test of Common Method Variance

Harman’s single-factor test for common method bias was employed in this investigation. Principal component analysis was performed using SPSS 26.0 software to extract principal components and examine the variance contribution of each component. The results in

Table 3. The rotated component matrix.

Constructs	Items	Factor Loadings
		1	2	3	4	5	6	7
Privacy Concerns	PC1	0.835
Privacy Concerns	PC2	0.841
Privacy Concerns	PC3	0.823
Cognitive Trust	CT1							0.729
Cognitive Trust	CT3							0.793
Cognitive Trust	CT4							0.665
Emotional Trust	ET1					0.776
Emotional Trust	ET2					0.764
Emotional Trust	ET3					0.681
Interactive Atmosphere	IA1		0.786
Interactive Atmosphere	IA2		0.745
Interactive Atmosphere	IA4		0.746
Scene Immersion	Sim2			0.715
Scene Immersion	Sim4			0.797
Scene Immersion	Sim5			0.831
Social Identity	SId3				0.769
Social Identity	SId4				0.735
Social Identity	SId5				0.741
Purchase Intention	PI1						0.723
Purchase Intention	PI3						0.813
Purchase Intention	PI4						0.719
Cumulative explained variance of the population (%)		74.524%

Note: the extraction method is principal component analysis; the rotation method is the varimax method; the explained variance ratio of the first factor is 11.58%.

show that the number of extracted factors is greater than 1, with a total of seven factors and a total explained variance of 75.25% (greater than 60%), indicating that the extracted factors can effectively explain the variability of the data. Moreover, the explained variance ratio of the first factor is 11.58% (less than <40%), which means that there is no single factor dominating the results; thus, the homogeneity deviation of the data in our study is relatively small.

4.2. Measurement Model Assessment

SPSS 26.0 is used to analyze the reliability and validity of the collected data and evaluate the measurement model. First, an Exploratory Factor Analysis (EFA) was performed using the principal component analysis method with Varimax rotation to analyze the factor structure and correlations between the scale items. The Kaiser-Meyer-Olkin (KMO) Measure of Sampling Adequacy yielded a value of 0.919, indicating that the sample size was adequate for factor analysis. The items were grouped into seven distinct factors (see Table 3), with factor loadings above 0.665 for each construct, suggesting good convergent validity. No significant cross-loadings were evident among the factors, supporting the convergent validity of the scales used. In addition, the average variance extracted (AVE) values for all constructs were greater than 0.566 in

Table 4. Reliability and validity.

Constructs	Items	Loadings	AVE	CR
Privacy Concerns Privacy Concerns Privacy Concerns	PC1	0.782	0.670	0.858
	PC2	0.895
	PC3	0.772
Cognitive Trust Cognitive Trust Cognitive Trust	CT1	0.792	0.566	0.796
	CT3	0.684
	CT4	0.776
Emotional Trust Emotional Trust Emotional Trust Interactive Atmosphere	ET1	0.759	0.584	0.808
	ET2	0.781
	ET3	0.752
	IA1	0.716
Interactive Atmosphere Interactive Atmosphere Scene Immersion	IA2	0.829	0.602	0.819
	IA4	0.778
	Sim2	0.777
Scene Immersion Scene Immersion Social Identity	Sim4	0.776	0.600	0.818
	Sim5	0.771
	SId3	0.784
Social Identity Social Identity Purchase Intention	SId4	0.763	0.582	0.807
	SId5	0.742
	PI1	0.814
Purchase Intention Purchase Intention	PI3	0.796	0.630	0.836
	PI4	0.770

Notes: CR = composite reliability; AVE = average variance extracted value.

, indicating good convergent validity. Both tests have confirmed high convergent validity.

The results of the reliability test show that the composite reliability (CR) values for all constructs in this survey were greater than 0.7, indicating high internal reliability (see Table 4). Lastly, the diagonal of the box represents the square root of the average variance extracted (AVE) for each construct in

Table 5. Discriminant validity.

	IA	PI	SID	SIM	ET	CT	PC
IA	0.794
PI	0.675	0.763
SID	0.661	0.655	0.775
SIM	0.611	0.549	0.600	0.776
ET	0.650	0.630	0.661	0.579	0.764
CT	0.651	0.654	0.686	0.576	0.661	0.752
PC	−0.422	−0.451	−0.453	−0.300	−0.545	−0.532	0.818

. In this case, the diagonal values are higher than the correlations outside of it, confirming that each variable measures a unique concept within the model and does not significantly overlap with others. These results validate the model’s structure, ensuring that the constructs are conceptually distinct.

4.3. Hypothesis Testing

The Structural Equation Modeling (SEM) method and AMOS tool were used to evaluate the basic structural model and the hypothesis about the effect of social scene characteristics in live streaming shopping, and the regression model was used to test the hypotheses of the moderating effect. AMOS-SEM is more effective and stable in dealing with measurement errors caused by variables [66], but it requires more sample sizes than PLS-SEM, and our data volume meets the requirements of a large sample size. Before hypothesis testing, it is necessary to evaluate and optimize the fitting degree of the research model, ensuring that the fitted model is valid and reliable for interpreting the relationships in the study. The model-fitting situation is shown in

Table 6. Model fit indices.

Fitting Index	χ2/df	GFI	AGFI	CFI	NFI	RMSEA	TLI
Recommended criteria	<3	>0.90	>0.90	>0.90	>0.90	<0.08	>0.90
Actual value	2.254	0.945	0.923	0.965	0.940	0.050	0.956

. All model fit indices have obtained the recommended values, which demonstrates a good fit with the research model.

Each hypothesis is then analyzed based on the estimated path coefficients and their significance levels (see Figure 2 and

Table 7. Hypothesis results.

Paths			Estimate	S.E.	C.R.	Significance
IA	→	ET	0.31	0.088	4.256	***
SIm	→	ET	0.17	0.055	2.779	0.005
SId	→	ET	0.36	0.059	5.138	***
IA	→	CT	0.29	0.088	4.256	***
SIm	→	CT	0.15	0.063	2.461	0.014
SId	→	CT	0.41	0.068	5.848	***
IA	→	PI	0.23	0.085	3.106	0.002
SIm	→	PI	0.06	0.056	0.981	0.326
SId	→	PI	0.20	0.070	2.468	0.014
ET	→	PI	0.17	0.073	2.322	0.020
CT	→	PI	0.21	0.068	2.807	0.005

Notes: IA = Interactive atmosphere; SIm = Scene immersion; SId = Social identity; PC = Privacy concerns, *** p < 0.001.

), incorporating mediating effects.

The results of the hypothesis analysis confirm that the social characteristics of the scene in live streams significantly influence purchase intention. The results show that the interactive atmosphere has a significant positive effect (β = 0.265, p = 0.002), reinforcing the importance of interactivity in consumer purchase decisions. H1a is supported. Likewise, social identity positively impacts purchase intention (β = 0.173, p = 0.014), indicating that consumers who feel part of a community are more likely to make purchases. H1b is supported. However, scene immersion does not have a significant direct effect on purchase intention (β = 0.055, p = 0.326). However, it can increase purchase intention by promoting trust, as customer trust plays a significant mediating role between itself and purchase intention. In Table 7, the test results of the mediating effect provide evidence for this. Therefore, considering both direct and indirect effects, scene immersion still has a significant impact on purchase intention, and H1c is supported.

Regarding consumer trust, both emotional trust H2a (β = 0.171, p = 0.020) and cognitive trust H2b (β = 0.190, p = 0.005) are confirmed to have positive effects on purchase intention, demonstrating that consumer trust, whether based on emotional connection or rational evaluation of the seller’s credibility, plays a pivotal role in the purchase decision. Finally, the interactive atmosphere has a positive impact on both emotional trust (β = 0.348, p < 0.001) and cognitive trust (β = 0.375, p < 0.001), indicating that an interactive scene in commerce service strengthens consumers’ emotional connection and rational perception of trustworthiness. Similarly, scene immersion positively affects emotional trust (β = 0.154, p = 0.005) and cognitive trust (β = 0.154, p = 0.014), Social identity shows the greatest impact on trust building, significantly influencing both emotional trust (β = 0.304, p < 0.001) and cognitive trust (β = 0.396, p < 0.001), suggesting that when consumers have a high identify with the community or social group in the live streaming, they perceive higher levels of trust.

We also conducted the mediation analysis using the Bias-Corrected Percentile method with a 95% confidence interval and a Bootstrap procedure with 2000 samples in AMOS to assess the effect of customer trust on the relationship between scene characteristics and purchase intention (see

Table 8. Mediation effect.

Research Hypothesis	Parameter	Bias-Corrected CI at 95%
		Estimate	Lower	Upper	p
H3a	IA→CT→PI	0.071	0.022	0.158	0.004
	IA→PI	0.265	0.076	0.463	0.009
	Total effect	0.336	0.147	0.535	0.001
H3b	IA→ET→PI	0.059	0.007	0.155	0.025
	IA→PI	0.265	0.076	0.463	0.009
	Total effect	0.324	0.143	0.509	0.001
H4a	SIM→CT→PI	0.029	0.001	0.088	0.035
	SIM→PI	0.055	−0.064	0.190	0.338
	Total effect	0.084	−0.035	0.223	0.150
H4b	SIM→ET→PI	0.026	0.003	0.076	0.020
	SIM→PI	0.055	−0.064	0.190	0.338
	Total effect	0.081	−0.037	0.213	0.159
H5a	SID→CT→PI	0.075	0.022	0.162	0.005
	SID→PI	0.173	0.013	0.332	0.034
	Total effect	0.248	0.116	0.402	0.000
H5b	SID→ET→PI	0.052	0.006	0.131	0.023
	SID→PI	0.173	0.013	0.332	0.034
	Total effect	0.225	0.066	0.389	0.004

Notes: IA = Interactive atmosphere. SIm = Scene immersion. SId = Social identity. ET = Emotional trust. CT = Cognitive trust. PI = Purchase intention.

). Overall, these findings reinforce the idea that trust plays an essential mediating role in converting live streaming scene characteristics into purchase intent. Customer trust plays a partial mediating role between interactive atmosphere and purchase intention, and between social identity and purchase intention. H3a, H3b, H4a, and H4b are supported. In contrast, customer trust plays a full mediating role between scene immersion and purchase intention; H5a and H5b are supported.

4.4. Moderating Effect

A hierarchical regression analysis was used to evaluate the moderation effect), following the approach employed by previous literature [67]. Moderation analysis reveals that privacy concerns play a significant role in the relationship between interactive atmosphere, scene immersion, social identity, and consumer trust, which has emotional and cognitive trust, although their effect varies depending on the independent variables. Specifically, H6a is supported, as privacy concerns were found to significantly and negatively moderate the relationship between interactive atmosphere and emotional trust (β = −0.138, p < 0.001). Likewise, H6b is supported, with a significant negative moderation effect observed between scene immersion and emotional trust (β = −0.104, p < 0.001). These findings suggest that while interactive features and immersive experiences generally foster emotional trust, their effectiveness diminishes when consumers harbor strong concerns regarding data privacy. Therefore, trust-building mechanisms rooted in emotional engagement may be less impactful among privacy-conscious users (see Figure 3). Conversely, the results do not support H6c, which hypothesized that privacy concerns would negatively moderate the relationship between social identity and emotional trust. The interaction term for this relationship was not statistically significant (β = −0.068, p > 0.05), indicating that the emotional trust derived from a shared sense of community and identity in the live streaming environment remains robust even in the presence of elevated privacy concerns. This finding implies that social belongingness may serve as a relatively stable driver of emotional trust, less susceptible to erosion by privacy-related apprehensions compared to interactive and immersive stimuli (see

Table 9. Hierarchical regression results for testing the moderating effect of privacy concerns on emotional trust.

Variable	Emotional Trust (ET)
	Model 1	Model 2	Model 3	Model 4	Model 5	Model 6
Control variables:
Age	0.147 ***	0.161 ***	0.156 ***	0.167 ***	0.159 ***	0.165 ***
Education	0.049	0.042	0.038	0.038	0.050	0.049
Employment	0.043	0.043	0.072	0.074	0.044	0.041
Independent variables:
IA	0.411 ***	0.416 ***
SIm			0.362 ***	0.354 ***
SId					0.413 **	0.411 ***
Moderating variables:
PC	−0.308 ***	−0.284 ***	−0.362 ***	−0.346 ***	−0.296 ***	−0.287 ***
Interaction term:
IA × PC		−0.138 ***
SIm × PC				−0.104 ***
SId × PC						−0.068
R2	0.390	0.409	0.364	0.374	0.389	0.393
Adjusted R2	0.384	0.402	0.358	0.367	0.383	0.386
F2	63.742 ***	57.248 ***	57.023 ***	49.593 ***	63.409 ***	53.732 ***

Notes: IA = Interactive atmosphere; SIm = Scene immersion; SId = Social identity; PC = Privacy concerns, f² = Effect Size of Path; R² = Coefficients of Determination, ** p < 0.01, *** p < 0.001.

).

Moreover, the hypothesized moderating effects of privacy concerns on the relationships between social scene characteristics; namely, interactive atmosphere, scene immersion, and social identity and cognitive trust, were not statistically supported. Specifically, H6d, H6e, and H6f did not achieve significance, as evidenced by the non-significant interaction terms (IA × PC: β = 0.014, p > 0.05) (SIm × PC: β = −0.062, p > 0.05) (SId × PC: β = 0.054, p > 0.05). These findings indicate that cognitive trust, which is grounded in analytical reasoning and the systematic evaluation of information, appears to be less susceptible to the influence of privacy concerns when compared to emotional trust (see

Table 10. Hierarchical regression results for testing the moderating effect of privacy concerns on cognitive trust.

Variable	Dependent Variable: Customer Trust (CT + ET)
	Model 1	Model 2	Model 3	Model 4	Model 5	Model 6
Control variables:
Age	−0.007	−0.009	0.002	0.008	0.003	−0.002
Education	−0.017	−0.016	−0.027	−0.027	−0.013	−0.013
Employment	−0.082 *	−0.082 *	−0.054	−0.052	−0.081	−0.078
Independent variables:
IA	0.417 ***	0.416 ***
SIm			0.368 ***	0.363 ***
SId					0.433 **	0.434 ***
Moderating variables:
PC	−0.295 ***	−0.297 ***	−0.394 ***	−0.340 ***	−0.277 ***	−0.283 ***
Interaction term:
IA × PC		0.014
Sim × PC				−0.062
SId × PC						0.054
R2	0.358	0.359	0.332	0.336	0.368	0.370
Adjusted R2	0.352	0.351	0.325	0.328	0.361	0.363
F2	55.637 ***	46.312 ***	49.513 ***	41.859 ***	57.875 ***	48.713 ***

Notes: IA = Interactive atmosphere; SIm = Scene immersion; SId = Social identity; PC = Privacy concerns, f² = Effect Size of Path; R² = Coefficients of Determination, * p < 0.05, ** p < 0.01, *** p < 0.001.

). One plausible explanation is that consumers engaging in cognitive trust formation may prioritize objective indicators, such as perceived competence, consistency, or transparency of the seller, over affective or contextual cues derived from the social environment. Thus, even when privacy concerns are elevated, consumers may still engage in rational processing that upholds their cognitive trust, rendering the moderating effect of privacy concerns negligible in this context.

5. Discussion

This study examines the characteristics of the social scene in live streaming commerce and empirically investigates the relationships between interactive atmosphere, scene immersion, and social identity, and their effects on emotional and cognitive trust, which in turn influence consumers’ purchase intention. Furthermore, it explores the moderating role of privacy concerns in these relationships. The results show that the social scene characteristics increase consumers’ purchasing intention, which aligns with the previous study by Qian et al. [8], Yin et al. [68], and Shang et al. [69], indicating that characteristics such as interactivity, entertainment, and social presence can have a positive impact on purchase intention. Aligned with studies related to consumers’ trust [49,70], this study confirms that consumer trust plays a critical role in driving purchase intention. Notably, this study has also found the mediating role of emotional trust and cognitive trust in the relationship between interactive atmosphere, scene immersion, social identity, and purchase intention. The results imply that the customers make their shopping decisions in live streaming largely depending on their emotional trust and cognitive trust in both the streamer and the platform. Interactive atmosphere, scene immersion, and social identity brought by live streaming scenes to the audience greatly meet the social needs of the audience’s psychology. As a result, emotional trust and cognitive trust are more driven by the social psychology of the live streaming scene that allows them to make predictions of purchasing decisions with a certain level of confidence. This further supports previous research that suggests there is a close relationship between customer trust and the live streaming environment [8,49].

Baker et al. [16] conceptualize service scenarios as comprising three dimensions: atmospheric elements, social elements, and design elements. Building upon previous work that emphasizes interactivity in live streaming [63,71], the present study contributes a comprehensive knowledge of the social characteristics of the live e-service scene and validates the role of three social scene elements based on customers’ cognition of meeting social needs. Therefore, our findings have increased our understanding of scene-building strategies in live e-service. Furthermore, our research findings confirm that this framework can effectively explain the impact of scenarios and trust on technology usage behavior, different from the previous explanation of technology usage behavior with this model, thus expanding the application of the cognition-affect-conation model in the live streaming commerce context.

In particular, the results also verify that privacy concern negatively moderates the relationships between the interactive atmosphere, scene immersion, social identity, and two types of customer trust which are cognitive and emotional. The stronger the privacy concern, the weaker the positive effects of interactive atmosphere and scene immersion on customer’s emotional trust. Accordingly, H6a and H6b are supported. These results support previous related studies on privacy concerns [63,72], which have proposed that users tend to be defensive about their privacy, and the defensive attitudes could have a significant impact on the customers’ decisions. However, H6c and H6f are not supported, indicating that privacy concerns do not significantly weaken the relationship between social identity and emotional trust, nor does it weaken the relationship between social identity and cognitive trust. One interpretation is that strong social identification in live streaming communities can override privacy concerns, as users perceive their identities to be validated and recognized. With the mutual infiltration of audience and anchor identities in live e-service, people no longer need to conceal their personal emotions and identities in virtual communication, and their opinions can be freely expressed. In such environments, trust emerges from shared values and emotional resonance, which diminishes the influence of privacy related hesitation.

In addition, the moderation effect of interactive atmosphere and scene immersion between cognitive trust by privacy concerns is not supported. A reasonable explanation can be obtained from the understanding of the mechanism of cognitive trust. Cognitive trust is built on the basis of individual rational evaluation. When people use reliable evidence as a criterion for judging whether to believe, cognitive trust will come into play [73,74]. This form of trust mainly comes from consumers’ awareness and understanding of the sociality of service scenarios. Once consumers recognize that the interactive atmosphere and scene immersion meet social needs, this cognitive trust will become deeply ingrained, and it is difficult to be influenced by the psychological tendency of privacy concerns. Unlike cognitive trust, emotional trust refers to trust established through emotional connections between people, which does not come from rational reasoning and understanding, but from feelings and sensation [75]. The psychological induction generated by the cognition of interactive atmosphere and immersion stimulates emotional trust, which is based on sensibility and is more susceptible to interference from personal psychological tendencies. Therefore, privacy concerns can generate emotional worries, thereby weakening the stimulating effect of the interactive atmosphere cognition and scene immersion cognition on emotional trust. Previous findings have revealed the direct impact of privacy concerns on trust, while our results reveal the interactive effects of privacy concerns and service scene elements on trust, making it more helpful for understanding the trust building mechanism and differences in the cultivation of two different types of trust in live e-service and e-commerce.

5.1. Theoretical Implications

The findings from this study enhance the theoretical understanding of the antecedents influencing consumers’ purchase intention in live streaming commerce, while also expanding the research perspective in this emerging domain. Previous literature has mainly studied some antecedents of live streaming shopping [9,38,52], such as physical features (e.g., situational normality and privacy security), marketing elements (e.g., competence of streamer, perception of price and risk, and interactivity), and psychological features of consumers (e.g., social presence, trust, and perceived enjoyment). However, limited attention has been paid to the role of social scene elements shaped by platforms and merchants within the live e-service environment. Addressing this gap, the present study investigates three core scene characteristics: interactive atmosphere, scene immersion, and social identity, from the perspective of consumer cognition, and examines how they shape purchase intention in live streaming commerce. In addition, while previous research focused on the influence of technical characteristics and background fitting of live streaming commerce on trust [52,69], this study sheds light on how consumer trust, both emotional and cognitive, mediates the influence of social scene characteristics on consumers’ purchase intentions. By doing so, the study identifies a cognitive mechanism of specific atmosphere, and a psychological mechanism based on trust that drives decision-making in live streaming commerce, offering a novel angle to understand purchase intention in this interactive and socially rich setting. This study also contributes theoretically by extending and deepening the application of the Cognition-Affection-Conation model in the context of live e-service. Our results validate the effectiveness of this framework in explaining user intentions in live streaming environments.

Moreover, the study makes an additional theoretical contribution by revealing the moderating role of privacy concerns in the relationship between social scene characteristics and consumer trust (emotional and cognitive). Although privacy concern is widely recognized as an important factor in live streaming shopping [49], few studies have explored the interacting effects of privacy concerns and social scene elements. This study focuses on viewer-related factors, e.g., privacy concern, and provides empirical evidence on the effect of privacy concern, emotional trust, cognitive trust, and three social scene characteristics on purchase intention in live streaming shopping. Thus, these findings offer new insights into how trust is constructed and influenced by both social and privacy-related factors in live streaming commerce. By focusing on viewer psychological-related factors such as privacy concerns and Cognitive-Affective mechanisms, the study advances our theoretical understanding of live streaming commerce purchase intention. It contributes to the live streaming marketing and commerce literature by highlighting how social scene cognition, rather than only physical or interactivity-driven factors, plays a crucial role in shaping trust and influencing purchase intention in this social and digital environment.

Furthermore, although this study is situated within the context of live streaming commerce, its findings offer broader theoretical implications for the e-commerce industry as a service-driven dominant. Live e-commerce today increasingly encompasses not just transactions, but also experiential, social, and emotional services elements that influence customer decision-making [76]. The identified roles of social scene characteristics, emotional trust, cognitive trust, and privacy concerns align closely with the service dominant logic in live service marketing, emphasizing the inspiration of intention through service scene cognition and psychological mechanisms. By examining how interactive environments and social identity foster trust and shape purchase intention, this research contributes to understanding how scene characteristics of live e-service, such as real-time engagement, satisfaction of social need, and trust-based interactions, can be effectively designed and managed. These insights are valuable for advancing theoretical discussion around the scene strategy of live e-service and trust building in real-time digital and social environments.

5.2. Practical Implications

Our study provides some managerial implications for firms that make an effort to attract consumers in live streaming sales campaigns. First, businesses should prioritize the cultivation of emotional atmospheres within live streaming services to address customer’s social needs and preferences in shopping, thereby strengthening customer’s trust. In terms of live streaming techniques, experienced hosts play a critical role. They must forge strong emotional bonds with viewers while offering timely, tactful responses to questions and feedback, encouraging proactive customer engagement. A key strategy involves leveraging experienced hosts who can establish emotional connections with viewers while providing timely and tactful responses to comments and inquiries. Furthermore, merchants should not only focus on delivering high-quality content to deepen sensory immersion and captivate audiences but also identify and incentivize viewers who contribute valuable content, activating their intrinsic motivation to participate. When these viewers receive recognition and engagement, they experience heightened excitement and deeper immersion in the live streaming. This immersive experience resonates among hosts, individual viewers, and the wider audience, initiating a virtuous cycle of engagement that strengthens consumer trust in the brand or platform. Finally, businesses, platforms, hosts, and customers must collaborate to create novel scenarios of social identity—spaces that alleviate stress and spark collective resonance. By shaping virtual environments that affirm shared identities and views during the live streaming, new scene cognitive frameworks emerge in customers’ mind, redefining customers’ communication behaviors and social interaction patterns.

Customer privacy concerns can weaken the positive impact of the live streaming social scene on customer trust. Platforms and merchants should try to avoid this situation. Platforms and merchants can enhance consumers’ awareness of “privacy policies” and “privacy protection tools” through pop-ups, short videos, and other forms of education. And platforms and merchants actively respond to users’ doubts and complaints and promptly address privacy issues related to social scenarios. At the same time, it is mandatory to require live streaming rooms to take privacy protection measures and remind users in prominent locations to carefully consider before sharing personal information, to reduce personal information leakage and abuse. In addition, companies can demonstrate their commitment and strength in data security and privacy protection through third-party authentication, user evaluations, and other means.

5.3. Managerial Implications

This study provides some managerial implications for firms that make an effort to attract consumers in live streaming sales campaigns. First, platform managers and content creators should prioritize the design of social live e-service scenes by fostering an interactive atmosphere and enhancing scene immersion through some measures, such as real-time comment interaction, dynamic visual design, and emotionally expressive streamers, to effectively build consumer trust. Emotionally expressive streamers must forge strong emotional bonds with viewers while offering timely, tactful responses to questions and feedback, encouraging proactive customer engagement. Hosts should continuously guide participants to voice their opinions and foster interactive and egalitarian dialogs within live streaming communities. These immersive and interactive elements not only capture attention but also fulfill users’ psychological and social engagement needs, encouraging deeper participation and stronger purchase intention.

Moreover, platforms, merchants, hosts, and customers must collaborate to create novel scenarios of social identity and scene immersion and let it become a space that alleviates stress and sparks collective resonance by encouraging the sharing of identities and views during live streaming. Platforms and merchants should not only focus on delivering high-quality content to deepen sensory immersion and captivate audiences but also identify and incentivize viewers who contribute valuable content, activating their intrinsic motivation to participate. Platforms and merchants should facilitate opportunities for users to form communities or engage with streamers and customers who share their values or lifestyles. Platforms and merchants also may provide participating customers with the opportunity to become a host and fully showcase themselves around the products being sold during the live broadcast, including cosplay performances and small theater shows. These measures help cultivate a sense of immersion and identification, which serve as a buffer against privacy concerns, as users who feel connected to a group or a personality may be more willing to participate and finally easy to be induced to purchase despite potential data risks. The formation of trusted social bonds within the streaming atmosphere plays a vital role in reducing resistance and fostering loyalty.

Finally, managing privacy proactively is crucial in maintaining long-term trust. While interactivity and immersion are powerful drivers of engagement, they must be balanced with transparent data practices. Providing clear privacy policies, an easy-to-understand permissions approach to privacy builds privacy protection credibility and reassures users, especially in an environment where personal interaction and data exchange are constant. Platforms and merchants should make an effort to enhance consumers’ awareness of “privacy policies” and “privacy protection tools” through continuous education by using pop-ups, short videos, and other forms of education. Platforms and merchants actively respond to users’ doubts and complaints and promptly address privacy issues related to social scenarios. At the same time, platforms need to make mandatory requirements for live streaming rooms to take privacy protection measures and remind users in prominent locations to carefully consider before sharing personal information, to reduce the risk of personal information leakage and abuse. In addition, companies can demonstrate their commitment and strength in data security and privacy protection through third-party authentication, user evaluations, and other means.

5.4. Limitations and Future Research Directions

Although this paper offers valuable contributions, it also has certain limitations that future research should address. First, despite efforts to include diverse regions and populations in the sample selection, achieving complete random sampling remains challenging. The sample for this study is primarily concentrated in Asia, with approximately half of the participants being students without work experience. Consequently, the data source inevitably reflects some degree of “convenience sampling”, which may introduce bias into the findings. Nevertheless, the demographic characteristics of the sample closely align with those of the broader population engaged in live streaming e-commerce, suggesting that the reliability of the conclusions is not substantially compromised. Future studies are encouraged to broaden the scope of data collection to encompass other regions and a wider range of occupational groups to enhance the generalizability of the results.

In addition, the study focuses on media competition, although previous research suggests that consumer’s perceptions of marketing warmth also influence purchase behavior. Future studies could explore the role of marketing warmth and how it affects consumer interaction and purchase decisions in live shopping. Second, this research examines the perception of price attractiveness and uncertainty as key psychological perceptions in live shopping. However, additional factors such as self-expression values (positive) and identity threats (negative) may also influence consumer decision-making. Future studies could investigate these psychological perceptions to further our understanding of consumer value evaluations in live shopping. Third, individual consumer differences, such as personality traits, shopping experience, and product knowledge, may moderate the effects of the marketing elements identified in this study. Future research should examine these interaction effects to gain a more nuanced understanding of consumer behavior in live streaming commerce. Furthermore, while this study focuses on streamer-related factors, particularly those that positively influence marketing and sales, viewer-related factors such as loneliness and fear of missing out may also significantly affect parasocial interactions and consumer-streamer identification. Future research could explore how these negative psychological factors influence consumer engagement, loyalty, and purchasing behavior. Another limitation is the platform-specific context, as this study focuses on TikTok and Douyin live streaming e-commerce. To enhance generalizability, future research should compare multiple platforms to determine whether emotional trust consistently acts as the dominant mediator across different ecosystems.

Moreover, this study relies on cross-sectional data, which captures relationships at a single point in time. Future research should employ longitudinal designs to examine how emotional and cognitive trust evolve with repeated exposure to live streaming content. Beyond trust, future research should explore cultural variations in consumer behavior across Western and Eastern markets. Furthermore, with the rise in AI-based live stream hosts and virtual influencers, it is crucial to examine how these non-human hosts compare to human hosts in building trust and driving engagement. Finally, emerging interactive elements, such as gamification, augmented reality (AR), and metaverse shopping, could transform live commerce experiences. Future research should evaluate their impact on consumer trust, engagement, and purchasing behavior. By addressing these limitations and exploring these directions, future research can contribute to a more complete understanding of consumer behavior and trust formation in live e-commerce.