Livestream e-commerce is one of the fastest-growing applications in e-commerce in recent years. In 2023, China’s online retail sales reached 15.42 trillion yuan, of which live-streaming e-commerce transactions totaled 4.9 trillion yuan, an increase of 40.48% compared with the same period last year. The number of active live-streaming users reached 540 million in 2023, up by 14.16% year-over-year. Among them, users aged 18–25 contributed 41.3% of platforms’ GMV, with an average monthly viewing time of 392 minutes (approximately 6.5 hours), a 14% increase from the previous year.

As the live-streaming e-commerce industry has rapidly expanded, some scholars have turned their attention to its academic underpinnings and issues. A CNKI annual cross-analysis of keywords shows that literature using the keyword “e-commerce live streaming” reached 252 articles in 2021, 280 in 2022, 256 in 2023, and 225 in 2024. An exact search for “e-commerce live streaming” on Google Scholar identified 3,070 related publications. A search of the Web of Science Core Collection using “e- commerce live streaming” yielded 406 results, with the three most frequent subject categories being “Business,” “Computer Science Information Systems,” and

“Management” (search date: February 5, 2025). These data indicate that research using “e-commerce live streaming” as a keyword is among current hotspots. However, most studies focus on business models, platform system optimization, and commercial organizations, with comparatively less attention to young people and student populations. A CNKI search using the keywords “e-commerce live streaming” and “college students” retrieved 218 related papers, while a Google Scholar search using “e-commerce live streaming” and “college students” yielded only 703 related publications.

Some studies indicate that in recent years, young people’s consumption values have shifted from traditional frugality to a pursuit of fashion, individuality, and quality. They place greater emphasis on spending for happiness, personal image, and environment, viewing consumption as a way to enjoy life rather than merely to meet basic survival needs (Wu, 2008). This shift reflects broader changes in social values. In the digital age in particular, consumption is tightly interwoven with social interaction and identity, exhibiting characteristics of “symbolic consumption”. Related research also finds that although young consumers are increasingly inclined toward rational planning (such as price comparison and group buying), issues like impulsive consumption and conspicuous consumption persist. Young people are susceptible to the influence of social media such as short videos and live streaming, which can lead to irrational purchasing behavior. This ambivalence highlights the coexistence of hedonism and rationality in young consumers’ decision-making. As a primary subset of the youth population, studying the impulsive buying behavior of college students in e-coomerce live strwaming can further enrich the research findings on youth consumer behavior, shed light on the evolution and formation of their consumption habits, and provide meaningful guidance for effectively fostering healthy consumption practices among college students.

Research Objectives

1. To examine how external stimuli—such as product factors, streamer factors, and live-streaming context factors—affect college students’ impulsive purchasing behavior.
2. To reveal the direct effects of personal psychological variables—such as self- control tendency, materialism tendency, and hedonic motivation—on impulsive consumption behavior.
3. To explore the mediating roles of self-control tendency, materialism tendency, and hedonic motivation in the pathways of external stimuli on impulsive consumption behavior.

The Stimulus–Organism–Response theoretical model (SOR model) is an important framework in psychology and behavioral science for explaining the mechanisms of individual behavior. Proposed by Mehrabian and Russell (1974) in the context of environmental psychology, the model posits that external stimuli (S) first trigger individuals’ internal cognitive or affective states (O), which in turn drive specific behaviors (R). It emphasizes that external stimuli influence behavior through the mediating role of the organism’s internal states (e.g., cognition, emotion), ultimately eliciting behavioral responses. The SOR model breaks through the mechanistic limitations of the earlier S–R (stimulus–response) paradigm by highlighting the dynamic mediating role of individual psychological processes. Su et al. (2019) found that consumers’ cognitive appraisal processes under emotional shocks significantly affect decision-making, with negative emotional appraisals often leading to consumption retreat. Wang and Aisihaer et al. (2022) used experimental methods to examine how streamers’ interactive content in live-stream marketing affects consumers’ purchase intentions and found that affective and cognitive responses play significant mediating roles in this process. Cognitive Appraisal Theory (CAT), as one of the core theories in affective psychology, emphasizes the decisive role of individuals’ cognitive appraisal of external events in the generation of emotions. First proposed by Lazarus and Folkman, its core view is that emotional responses arise from individuals’ dual appraisal of the degree to which external stimuli align with their goals and needs. Feng and Qi (2014), drawing on systemic functional semiotics, extended the application of CAT to multimodal interactive contexts, constructing a three-layer model comprising eliciting conditions, internal attitudes, and attitude expression to systematically analyze the mechanisms by which multimodal information contributes to the formation and transformation of attitudes.

Within the SOR framework, product factors, streamer factors, and live-streaming context factors serve as external stimuli that trigger consumers’ instant purchase behavior, while consumers’ cognitive appraisals and emotional states play mediating roles. In terms of product factors, this is reflected specifically in the degree of match between price and product functionality and quality—that is, price sensitivity or the product’s value for money—as well as in pricing discounts and promotional policies. When consumers perceive a high degree of alignment between a product’s price and its perceived value, their purchase intention tends to increase accordingly; conversely, if pricing exceeds consumers’ reasonable expectations of the product’s intrinsic value, perceived value may decrease, thereby inhibiting the formation of purchase decisions (Gong & Jiang, 2023). Promotional policies not only reduce consumers’ perceived risk during the decision-making process but also, by creating a sense of urgency to buy within a short time window, further strengthen consumers’ impulse to purchase immediately.

In terms of streamer-related factors, the key aspects include the reliability and professionalism of the streamer, as well as the creation of social presence. For sales- oriented streamers, reliability (e.g., integrity, public praise) and professionalism (e.g., product knowledge, accuracy of recommendations) are regarded as critical components that foster consumer trust (Truong, 2024). For entertainment-oriented streamers, physical attractiveness and entertainment performance (e.g., sense of humor, talent displays) are core attributes that satisfy viewers’ hedonic needs and, to some extent, indirectly encourage purchasing behavior (Liu & Shi, 2020). Skill-based streamers (e.g., beauty tutorial hosts) rely on superior skill demonstrations during livestreams (such as makeup techniques and demonstrations of product usage) to convey the functional value of products, thereby directly influencing purchase intention (Daisy Lee & Wan,C. 2023). By cultivating a high level of social presence during livestreams, streamers can also positively influence consumer behavior by providing emotional and participatory experiences, such as instant interaction, responding to live comments, and personalized answering (Sawmong, S. 2022).

In terms of livestream context factors, stimuli are reflected in the authenticity of content, interface design, and immersion. High-quality visual presentation (e.g., multi- angle shooting, real-time effect demonstrations) can effectively enhance consumer enjoyment, thereby promoting impulse purchases (Yang, Cao et al., 2022). Apparel products provide more direct visual stimuli and immersive experiences for consumers through dynamic try-ons or runway shows (Li, Wang et al., 2022). The mediating effect of interface design on impulse buying can reach as high as 37%; concise, clear visual information and user-centered interaction design often heighten arousal, making consumers more absorbed in the shopping context. High levels of immersion typically encompass multiple dimensions such as perceived control, interest, attention, and curiosity. When consumers experience strong immersion on livestreaming platforms, their engagement in the shopping process increases accordingly, significantly boosting the likelihood of impulse purchases or immediate decisions (Cui, Liu et al., 2022).

Research on consumer behavior mainly covers impulse buying and perceived value and risk uncertainty. In the livestream e-commerce context, consumers systematically process the product information conveyed by streamers and gradually form perceptions of product value, which directly influence their purchase intentions. Perceived value typically includes three dimensions: practical value, hedonic value, and social value. Liu, Meng, & Chen et al. (2020) point out that perceived trust is often jointly constructed by the streamer’s professionalism and the authenticity of product information, and is reinforced through long-term interaction; it usually plays a central role in purchase behavior. A streamer’s professionalism and interactive performance are directly linked to consumers’ dual perceptions of practical and hedonic value, thereby affecting purchase intention. Li and Zheng (2024) indicate that practical value has a significant and direct positive effect on purchase intention, while hedonic and social value mainly promote purchase intention indirectly by enhancing consumers’ trust in the platform and streamer as well as their emotional experience. Consumers with higher levels of trust are more likely to experience purchase impulses when faced with product recommendations, an effect that is especially pronounced in impulse-buying scenarios (Li, Jiang et al., 2022).

Research Hypothesis

Based on the above literature review, this paper formulates its research hypotheses. All hypotheses are presented as “Hypothesis (H) + number,” with the expected direction of effect (+/−) indicated in parentheses.

• H1: The higher the perceived value for money, scarcity, and uniqueness of a product, the stronger college students’ impulse buying tendency (+).
• H2: The higher the streamer’s professional level, physical appearance, reputation, and brand recognition, the stronger the impulse buying tendency (+).
• H3: The better the live-stream’s set design and atmosphere, and the more active the real-time interaction, the stronger the impulse buying tendency (+).
• H4: The stronger the self-control tendency, the lower the impulse consumption tendency (−).
• H5: Self-control plays a negative mediating role between external stimuli and impulsive purchasing (−).
• H6: The stronger the materialism tendency, the higher the impulse consumption tendency (+).
• H7: Materialism plays a positive mediating role between external stimuli and impulsive purchasing (+).
• H8: The stronger the hedonic motivation tendency, the higher the impulse consumption tendency (+).
• H9: Hedonic motivation plays a positive mediating role between external stimuli and impulsive purchasing (+).

Conceptual Framework

Based on the SOR theoretical framework and cognitive appraisal theory, this study selects product factors, streamer factors, and live-streaming context factors as external stimuli, and adopts self-control tendency, materialism tendency, and hedonic motivation as factors in self-cognition and appraisal. The conceptual framework of this study is shown in Figure 1.

Figure 1. Conceptual framework Source: Author compilation Research Methodology

All latent variables in this study were drawn from mature scales in highly cited Chinese and English literature. We adopted the internationally recognized “translation – back-translation–expert review” procedure to ensure semantic equivalence and used a 5-point Likert scale for measurement. For context-adapted items tailored to the live- streaming setting, three scholars in e-commerce and two veteran streamers were invited to assess content validity.

To ensure semantic fit and measurement quality of the scales, a small-scale pilot study was conducted prior to the main large-scale survey. The pilot sample was drawn           from one comprehensive university in the eastern region and one in the central region. A total of 130 questionnaires were collected; after removing extreme data and invalid responses, 110 valid samples were retained. In the reliability and validity tests of the pilot scale, the KMO value was 0.906, and Bartlett’s test of sphericity was significant (χ² = 8163.994, df = 666, p < 0.001), supporting factor analysis. The Cronbach’s α for each variable ranged from 0.88 to 0.89, and the composite reliability (CR = 0.89) exceeded 0.85, indicating good reliability.

Based on a stratified cluster random sampling framework, the authors collected data via an online self-administered questionnaire from 12 universities across the eastern, central, and western regions. The survey’s front-end logic included CAPTCHA, reverse-scored items, and an average response-time threshold (responses < 200 seconds or > 3,000 seconds were automatically excluded). Single IP and same-device submission limits were enforced, and the survey window period lasted two weeks. The study adhered to procedures for anonymity, informed consent, and ethics review. In total, 695 raw questionnaires were collected; after removing duplicates, logically inconsistent responses, and outliers, 604 valid samples were retained.

Using SPSS 25.0 for reliability and validity testing, the scale’s Cronbach’s α was 0.890, exceeding 0.85 and indicating high data reliability. For validity testing, the KMO value was 0.971, greater than 0.8, demonstrating excellent data validity and strong suitability for factor analysis.

Table 1 Cronbach Reliability Analysis

Source: Data and information from this research Table 2 KMO and Bartlett's Test of Sphericity

Source: Data and information from this research

Data Analysis

Descriptive Statistical Analysis

The survey results show that among 604 respondents, 466 were female, accounting for 77.15%, and 138 were male, accounting for 22.84%. There were 254 freshmen (42.05%), 141 sophomores (23.34%), 105 juniors (17.38%), 62 seniors (10.26%), and 42 postgraduate students (6.95%). Among the respondents, 54 had a monthly disposable allowance under 1,000 Chinese Yuanc(CNY) (8.94%); 496 had 1,001–2,000 CNY (82.11%); 41 had 2,001–3,000 CNY (6.78%); 7 had 3,001–4,000 CNY (1.15%); and 6 had over 4,001 CNY (0.99%). Among the respondents, 350 watched e-commerce livestreams 0 (exclusive)–1 hour per week (57.94%); 167 watched 2–4 hours (27.65%); 51 watched 5–10 hours (8.44%); and 36 watched more than 10 hours (5.96%). It should be noted that this survey excluded data from those who never watch e-commerce livestreams (or whose weekly viewing time is zero), which accounted for 18.05%. Among the respondents, 354 made purchases via e-commerce livestreams 1–2 times per month (58.61%); 185 made purchases 3–5 times (30.63%); 39 made purchases 6–9 times (6.45%); and 26 made purchases more than 10 times (4.30%). It should be noted that this survey excluded data from those who never shop via e-commerce livestreams (or whose monthly purchase count is zero), which accounted for 53.03%.

Structural Equation Modeling Analysis

Based on the model fit indices, the model demonstrates good performance in fitting the observed data, making it suitable for structural equation modeling analysis. First, the chi-square to degrees of freedom ratio (χ²/df = 2.99) is below the recommended threshold of 3; the GFI is 0.903, exceeding the ideal benchmark of 0.90; and the RMSEA is 0.059, below the recommended value of 0.10, indicating an excellent fit. In addition, both the CFI (0.907) and IFI (0.908) meet the criterion of being greater than 0.9, and both the PGFI (0.714) and PNFI (0.742) exceed the threshold of 0.5.

Table 3 Model Fit Indices

Source: Data and information from this research

Based on the structural equation modeling (SEM) results, the relationships among variables and their path coefficients reveal multiple factors influencing consumer behavior. Specifically, the factor loadings of the four observed variables for Product Factors (S-P) are S-P4 (0.714), S-P3 (0.612), S-P2 (0.642), and S-P1 (0.475), highlighting the importance of product novelty, brand reputation, scarcity, and price value. Similarly, the factor loadings of the four observed variables for Host Factors (S- H) are S-H4 (0.615), S-H3 (0.672), S-H2 (0.664), and S-H1 (0.726), indicating the significant roles of anchor reputation, anchor authenticity, anchor attractiveness, and anchor expertise. The four observed variables for Livestream Context Factors (S-L) have factor loadings of S-L4 (0.646), S-L3 (0.800), S-L2 (0.825), and S-L1 (0.796), underscoring the importance of information overload, interactivity, social proof, and scene atmosphere.

The path coefficient between Product Factors (S-P) and Self-Control (O-SC) is −0.205, indicating a negative effect, which suggests the inhibitory role of Self-Control on consumer behavior. The path coefficients between Product Factors (S-P) and Materialism (O-MA) and Hedonic Motivation (O-HM) are 0.615 and 0.475, respectively, with p-values less than 0.05, indicating significant positive relationships. The path coefficient between Host Factors (S-H) and Self-Control (O-SC) is 0.443, with a p-value less than 0.05, indicating a significant positive relationship. The path coefficient between Livestream Context Factors (S-L) and Self-Control (O-SC) is −0.512, with a p-value of 0, indicating a significant negative relationship, suggesting a pronounced inhibitory effect of Self-Control on it. Meanwhile, the path coefficient between Livestream Context Factors (S-L) and Hedonic Motivation (O-HM) is 0.679, with a p-value of 0, indicating a significant positive relationship.

The path coefficient from Self-Control (O-SC) to Impulse Buying (R-IB) is −0.259, with a p-value of 0, indicating a significant negative effect of Self-Control (O-SC) on Impulse Buying (R-IB). The path coefficient from Materialism (O-MA) to Impulse Buying (R-IB) is 0.321, with a p-value of 0, indicating a significant positive effect of Materialism (O-MA) on Impulse Buying (R-IB). The path coefficient from Hedonic Motivation (O-HM) to Impulse Buying (R-IB) is 0.39, with a p-value of 0, indicating a significant positive effect of Hedonic Motivation (O-HM) on Impulse Buying (R-IB).

Table 4 Model Regression Coefficients Summary Table

Source: Data and information from this research

Figure 2. Structural Equation Model Results Diagram Source: Data and information from this research

Mediation Effect Testing

To further examine the mediating roles of Self-Control (O-SC), Materialism (O- MA), and Hedonic Motivation (O-HM) in the effects of Product Factors (S-P), Host Factors (S-H), and Livestream Context Factors (S-L) on Impulse Buying (R-IB), this study conducted mediation effect tests and obtained the following results.

Table 5 Mediation Analysis Results

*p<0.05 **p<0.01

Source: Data and information from this research

From the table above, the mediation analysis involves five models, as follows:" R-IB=0.869+0.191*S-P+0.085*S-H+0.326*S-L O-SC=4.071+0.035*S-P+0.036*S-H-0.215*S-L O-MA=1.787+0.234*S-P+0.135*S-H+0.114*S-L  O-HM=0.901+0.131*S-P+0.117*S-H+0.288*S-L R-IB=1.327+0.125*S-P+0.040*S-H+0.176*S-L-0.247*O-SC+0.170*O- MA+0.270*O-HM

Table 6 Summary Table of Mediation Analysis Results

Source: Data and information from this research

The mediation analysis results indicate that along the paths “S-P => O-SC => R- IB” and “S-H => O-SC => R-IB,” the mediating effect of O-SC is not significant; along the paths “S-P => O-MA => R-IB” and “S-L => O-MA => R-IB,” O-MA exhibits a partial mediating effect; along the paths “S-P => O-HM => R-IB” and “S-L => O-HM => R-IB,” O-HM exhibits a partial mediating effect; along the path “S-L => O-SC => R-IB,” O-SC exhibits a partial mediating effect; along the path “S-H => O-MA => R- IB,” O-MA exhibits a full mediating effect; and along the path “S-H => O-HM => R- IB,” O-HM exhibits a full mediating effect.

Based on the above structural equation modeling analysis and the mediation effect test, the hypothesized test results of this study are as follows.

Table 7 Summary of Hypothesis Testing

Source: Data and information from this research

Based on the research objectives, the following conclusions can be drawn:

Accepted

The cohort of college students participating in e-commerce livestream consumption tends to be in lower grade (freshmen account for 42.05%), with a significantly higher proportion of female than male. Among students who engage with e-commerce livestreams, 82.11% have a monthly discretionary budget of 1,001 –2,000 CNY. The survey finds that 18.05% of students almost never watch e-commerce livestream videos each week, and 53.03% of students almost never make purchases during livestreams.

In terms of products, product novelty satisfies functional value while also activating emotional value, enabling rapid purchase decisions through heuristic processing. If the streamer further reinforces these quality signals with professional explanations and authoritative endorsements, rational and emotional dual pathways can be stacked in a short time, significantly amplifying purchase motivation. Meanwhile, studio design, lighting, and background music in the livestream environment create a highly immersive atmosphere, inducing a “flow” state characterized by time distortion and diminished self-awareness. Dopamine-driven anticipatory pleasure rises rapidly, and, when compounded by social conformity pressures from bullet comments, flash - sale prompts, and cues of inventory scarcity, college students—whose cognitive control is not yet fully developed—become nearly defenseless.

Impulse buying by college students in e-commerce livestream contexts is jointly influenced by three interacting psychological forces. Self-control affects impulsive purchases primarily by suppressing materialism, exerting an indirect effect, while materialism and hedonic motivation serve as two major positive drivers representing

value orientation and emotional arousal, respectively. The highly time-sensitive social atmosphere and catalytic environment with conspicuous symbols cultivated by e- commerce livestreams further amplify the latter two forces, placing greater strain on cognitive inhibition systems.

Recommendation

1. Consumption Advice for College Students

College students should establish actionable, trackable, and iterative behavior scheme across multiple dimensions to ensure a healthy and rational approach to livestream e-commerce consumption. First, college students should enhance awareness of actual product needs by preparing a shopping list in advance and assessing necessity. Second, when watching livestreams, they should practice emotional regulation, identify scarcity cues and social proof embedded in sales pitches, and avoid using shopping to cope with stress as an emotional substitute; also limit the time and frequency spent viewing livestreams. Third, it is necessary to set strict budgets for living expenses, and manage fixed expenses, necessary variable expenses, and discretionary spending in separate accounts; consider establishing a “consumption contract” with family or peers, maintaining a shared ledger, and holding regular “financial review meetings.” Finally, they should also strengthen rational thinking skills and improve quality of life in the real world; allocate resources with a long-term value orientation to health, skills, and relationships; establish a correct view of material possessions and standards for sustainable consumption so that consumption returns to its essential role of serving life.

2. Recommendations for Merchants For merchants

For merchants, the synergy between product novelty and symbolic value in product development should be prioritized. Cutting-edge materials and innovative processes can be applied to spark a sense of functional newness, while building a symbolic system through cultural narratives, identity symbols, and aesthetic styles to achieve differentiated premiums and cross-category expansion amid homogenized competition. In allocating streamers, to dynamically develop a dual-axis approach of professionalism and approachability with dynamic development is essential. The former builds trust through professional knowledge and verifiable use cases; the latter enhances stickiness through natural interaction, emotional resonance, and value alignment, which also combines structured scripts, contextual demonstrations, and measured improvisation to balance authoritative endorsement with human warmth. Merchants should treat emotion as a core variable in livestream scene design and implement systematic emotional modulation: tp use lighting, color, musical tempo, camera language, and chat moderation to jointly control the arousal–pleasure curve, reducing cognitive load while improving conversion rates. At the same time, they can actively advance social responsibility and value co-creation by bringing ESG topics—such as green supply chains, sustainable packaging, empowerment of vulnerable groups, and community co- benefits—into product selection and content narratives from the outset, and establish long-term trust and compounding brand goodwill through transparent goals and measurable metrics.

3. Recommendations for Livestreaming Platforms

For livestreaming platforms, rational elements should be systematically integrated into core algorithms and interface design. By optimizing recommendation logic, they should reduce the proportion of high-stimulating, emotion-driven content and increase the proportion of signals based on product quality and after-sales reputation, and introduce rational prompts in the front-end interface and de-emphasize impulse triggers such as countdown timers and “lowest price on the internet.” Streamers should build their personal brands around professionalism and self-control, establish verifiable credentials and product selection standards, provide transparent information sources and real-world test evidence, avoid exaggeration and absolute claims, disclose financial relationships and advertising attributes, and respect consumers’ rights to compare and hesitate, replacing short-term conversions with long-term trust. In addition, merchants, platforms, and streamers should jointly build collaborative mechanisms that encourage rational consumption, increasing the emphasis on low refund rates and low complaint rates, incorporating long-term compliance indicators—such as after-sales experience and product/information integrity—into KPIs; and establishing penalties for false advertising, excessive marketing, and abnormal returns, to ultimately achieve a sustainable balance between efficiency and responsibility.

1. Cui, Y., Liu, Y., Yu, C., and Liu, X. "Investigating the Key Drivers of Impulsive Buying Behavior in Live Streaming." Journal of Global Information Management, vol. 30, no. 6, 2022, pp. 1–23.
2. Lee, Daisy, and C. Wan. "The Impact of Mukbang Live Streaming Commerce on Consumers’ Overconsumption Behavior." Journal of Interactive Marketing, 2023.
3. Feng, D. Z., and Y. J. Qi. "态度意义的多模态建构——基于认知评价理论的分析模式 [The Multimodal Construction of Attitudinal Meaning: An Analytical Model Based on Cognitive Appraisal Theory]." 现代外语 (Modern Foreign Languages), 2014.
4. Gong, X. X., and X. T. Jiang. "Understanding Consumer Impulse Buying in Livestreaming Commerce: The Product Involvement Perspective." Frontiers in Psychology, 2023.
5. Li, M. W., Q. J. Wang, et al. "Understanding Consumer Online Impulse Buying in Live Streaming E-Commerce: A Stimulus-Organism-Response Framework." International Journal of Environmental Research and Public Health, vol. 19, no. 12, 2022, p. 7312.
6. Li, Z., and S. Zheng. "The Influence of the Language Style of the Anchor on Consumers’ Purchase Intention." Frontiers in Psychology, 2024.
7. Li, G. M., Y. Jiang, et al. "The Influence Mechanism of Interaction Quality in Live Streaming Shopping on Consumers’ Impulsive Purchase Intention." Frontiers in Psychology, 2022.
8. Liu, F. J., L. Meng, S. Y. Chen, et al. "网红直播对消费者购买意愿的影响及其机制研究 [Influence of Internet Celebrity Live Streaming on Consumers’ Purchase Intention and Its Mechanism]." 管理学报 (Chinese Journal of Management), vol. 17, no. 5, 2020, pp. 717–724.
9. Liu, P. S., and Y. D. Shi. "直播带货营销模式对消费者购买决策的影响机制 [Mechanism of Live Commerce Marketing Model on Consumers’ Purchase Decisions]." 中国流通经济 (China Business and Market), vol. 34, no. 6, 2020, pp. 120–131.
10. Mehrabian, A., and J. A. Russell. An Approach to Environmental Psychology. MIT Press, 1974.
11. Sawmong, S. "Examining the Key Factors That Drive Live Stream Shopping Behavior." Emerging Science Journal, 2022.
12. Su, L. J., S. Huang, et al. "Toward a Model of Destination Resident–Environment Relationship: The Case of Gulangyu, China." Journal of Travel & Tourism Marketing, 2019.
13. Truong, T. D. "Factors Influencing Gen Z’s Purchasing Intention and Behaviour in Livestreaming E-Commerce." International Journal of Social Science and Economic Research, vol. 9, no. 1, 2024, pp. 1–14.
14. Yang, J. J., C. Cao, et al. "Effects of Interface Design and Live Atmosphere on Consumers’ Impulse-Buying Behaviour from the Perspective of Human–Computer Interaction." Sustainability, vol. 14, no. 23, 2022, p. 16222.
15. Wang, X. L., N. Aisihaer, et al. "Research on the Impact of Live Streaming Marketing by Online Influencers on Consumer Purchasing Intentions." Frontiers in Psychology, 2022.
16. Wu, C. P. "改革开放 30 年与青年消费观念的变迁 [Changes in Youth Consumption Concepts Over 30 Years of Reform and Opening-Up]." 中国青年研究 (China Youth Research), 2008.