Next Article in Journal
Adoption of Incremental and Radical Innovations in E-Commerce Delivery: Evidence from Smart Lockers and Autonomous Drones Using the UTAUT Model
Previous Article in Journal
Designing a Sustainable Supply Chain Network for Perishable Products Integrating Internet of Things and Mixed Fleets
 
 
Search for Articles:
Title / Keyword
Author / Affiliation / Email
Journal
Article Type
 
 
Advanced Search
 
Section
Special Issue
Volume
Issue
Number
Page
 
Journals
JTAER
Volume 20
Issue 2
10.3390/jtaer20020138
jtaer-logo
Submit to this Journal Review for this Journal Propose a Special Issue
► Article Menu

Article Menu

share Share announcement Help format_quote Cite question_answer Discuss in SciProfiles
first_page
settings
Order Article Reprints
Font Type:
Arial Georgia Verdana
Font Size:
Aa Aa Aa
Line Spacing:
Column Width:
Background:
Article

From Certainty to Doubt: The Impact of Streamer Expression Certainty on Consumer Purchase Behavior in Live-Stream E-Commerce

by
Yinjiao Chen
1,2,*,
Qianqian Han
3,
Zhihua Lian
4,
Weiming Zhu
5 and
Yushi Jiang
6,*
1
School of Management, Guangdong University of Science and Technology, Dongguan 523083, China
2
School of Business Administration, Macau Millennium College, Macau, China
3
School of Business, Nanjing Audit University, Nanjing 211815, China
4
School of International Business, Xiamen University Tan Kah Kee College, Zhangzhou 363105, China
5
School of Business, Macau University of Science and Technology, Macau, China
6
School of Economics and Management, Southwest Jiaotong University, Chengdu 611756, China
*
Authors to whom correspondence should be addressed.
J. Theor. Appl. Electron. Commer. Res. 2025, 20(2), 138; https://doi.org/10.3390/jtaer20020138
Submission received: 17 March 2025 / Revised: 29 May 2025 / Accepted: 30 May 2025 / Published: 9 June 2025
(This article belongs to the Topic Interactive Marketing in the Digital Era)
Browse Figures
Versions Notes

Abstract

This study explores the impact of streamer expression certainty on consumer purchase behavior in live-stream e-commerce and its underlying mechanisms. Based on an empirical analysis of 3842 product live-stream data from the Douyin platform and an experimental study, the results show that streamer expression certainty has a significant inverted U-shaped relationship with consumer purchase behavior. Moderate expression certainty enhances consumers’ purchase intention, while excessively high or low expression certainty weakens its effect. Furthermore, perceived competence and perceived warmth mediate this process. Moderate expression certainty enhances consumers’ perceptions of the streamer’s professionalism and sincerity, thereby promoting purchase behavior. However, overly high expression certainty may reduce perceived warmth, leading to decreased consumer trust. This study further reveals the key role of language style in live-stream e-commerce and provides practical insights for optimizing streamers’ language strategies and improving live-stream effectiveness. The study’s limitations and directions for future research are also discussed.

1. Introduction

With the rapid development of the digital economy, live-stream e-commerce has become an important channel for consumers to obtain product information and make purchase decisions [1,2]. In this field, streamers, as the core figures for information transmission, are not only product recommenders but also connectors between consumers and brands [3,4]. By generating product-related content and interacting with consumers in real-time during live-streams, streamers can quickly capture consumer attention, convey product value, and stimulate purchase behavior [5,6]. According to statistics, the scale of China’s live-stream e-commerce market exceeded CNY 4.9 trillion in 2023, making it one of the main drivers of retail growth [7,8]. Therefore, companies are increasingly focusing on engaging with consumers through innovative interactive approaches to build or strengthen their brand image [9]. However, the success of live-stream e-commerce relies not only on the products themselves, but more importantly on how streamers build trust with consumers through effective language expression [10]. As competition intensifies, the role of streamers’ expression styles in attracting consumer attention, shaping a trustworthy image, and promoting purchase behavior is becoming increasingly important [11,12]. Therefore, how to optimize streamers’ language expression styles to enhance consumer trust and purchase intention has become a key issue in the field of live-stream e-commerce.
In practice, expression certainty is a language strategy frequently used by streamers, referring to the way individuals convey their high confidence and certainty about the information through language [13,14]. Most live-stream streamers hope to boost viewers’ confidence through such a tone, aiming to influence consumers’ purchase decisions [15]. Moreover, the majority of influencers and companies believe that using certainty expressions in their language helps them appear more credible. As a result, they frequently adopt certainty-based language strategies in livestreaming [13]. However, expression certainty does not always yield positive results [16]. On the one hand, in the field of consumer behavior, expression certainty (as opposed to uncertainty) is expected to convey the communicator’s confidence and professionalism, thereby attracting consumers’ interest [17]. On the other hand, research by Moy et al. (2024) in the context of crowdfunding found an inverted U-shaped relationship between expression certainty and crowdfunding success [18]. This duality makes the effectiveness of expression certainty in practice uncertain. Despite its widespread use in live-stream practices, existing research seldom explores its specific impact on consumer purchase behavior. In particular, in the context of live-stream e-commerce, it remains unclear how expression certainty influences consumer purchase decisions and its underlying mechanisms.
To fill this research gap, this paper intends to explore the impact of streamers’ expression certainty on consumer purchase behavior and its underlying mechanisms in live-stream e-commerce, based on stereotype theory. Stereotype theory suggests that people infer others’ competence and warmth through their language expressions, and form overall impressions and attitudes toward them based on these inferences [19]. For streamers, consumers’ perceptions of the streamer’s competence and warmth directly influence their trust in the streamer and purchase intention [13,20]. In live-stream e-commerce, streamers aim to present an image of both competence and warmth through their language to enhance consumer trust and drive purchases. However, stereotype theory also emphasizes a compensatory effect between competence and warmth; when an individual exhibits higher competence, their perceived warmth often decreases, and vice versa [21,22]. Based on this logic, we propose that there may be an inverted U-shaped relationship between streamers’ expression certainty and consumer purchase behavior. When expression certainty is moderate, consumers perceive the streamer as highly competent, thereby increasing trust and promoting purchase behavior. However, when expression certainty is too high, consumers’ perceived warmth of the streamer decreases, which may weaken trust and inhibit purchase behavior.
This study aims to explore the impact mechanism of streamers’ expression certainty on consumer purchase behavior through a multi-method validation process to fill this research gap. Specifically, this research analyzes the live-stream content of 242 sessions on the Douyin platform, covering 3842 products and 4.96 million unstructured text data. Additionally, a behavioral experiment is conducted to comprehensively examine the influence of streamers’ expression certainty on consumer purchase behavior and its underlying mechanisms.
This study makes significant contributions both theoretically and practically. First, it introduces expression certainty into the live-stream e-commerce domain for the first time, filling the gap in the existing literature on the language expression styles of streamers and enriching the application scenarios of language persuasion and stereotype theory [23]. Second, the study reveals the inverted U-shaped relationship between streamers’ expression certainty and consumer purchase behavior, as well as its underlying mechanisms, providing theoretical support for optimizing streamers’ language expression strategies in live-stream e-commerce. The research not only contributes to the literature on live-stream e-commerce, language persuasion, and stereotype theory but also offers practical suggestions for companies and streamers on how to adjust expression certainty to enhance marketing effectiveness in live-stream marketing.

2. Theoretical Background and Hypotheses Development

2.1. Language Strategies in Live-Stream E-Commerce

Language strategy is considered as the use of certain functional words (such as pronouns, prepositions, etc.) by communicators to express the same content in different forms [24]. Influencers build trust with customers through clever language strategies [25], stimulate purchase desire [26], and drive sales performance [27]. Therefore, language strategy is regarded as an important topic in e-commerce [28]. Due to the different ways words are used and information is presented, different language strategies can elicit various responses from information receivers [27].
Research on language strategies used by streamers in live-stream e-commerce can be roughly divided into two aspects. On one hand, it focuses on the classification of language strategies, namely task-oriented language and social-oriented language [29]. Task-oriented language aims to provide more detailed information to improve efficiency [30]. Social-oriented language is considered to enhance psychological closeness through social interactions, which is more conducive to building trust [31]. Research shows that socially oriented language, rather than task-oriented language, produces more effective customer behavior [30]. On the other hand, the focus is on the influencing factors of the streamer’s language strategies. Firstly, Yang and Wang (2022) conducted a keyword analysis based on a corpus and compared the differences in vocabulary usage between male and female streamers. They found significant gender differences in the selection and pragmatics of address terms [32]. Secondly, different language strategies can trigger different consumer responses, which are closely related to the product type [27]. For search products task-oriented language is more appropriate, as it provides specific and functional information that meets consumers’ needs and improves information processing efficiency and utilitarian value [4]. However, when purchasing experience products online, consumers are usually unable to fully evaluate the product’s functions based on available information [33]. Therefore, the role of social language from the streamer seems to be more effective. By making the streamer appear empathetic and enhancing the experiential value of the product, social language deepens consumers’ understanding of the product and increases social influence during the live-stream session [34].
In conclusion, streamer language strategies play a crucial role in shaping consumer behavior and optimizing live-stream e-commerce performance. Existing studies remain limited in scope, lacking a comprehensive investigation into other critical linguistic strategies. Given that language is a fundamental tool for persuasion and engagement in live-stream contexts, expanding research to encompass more diverse linguistic strategies—including expression certainty and other discourse techniques—would provide a more holistic understanding of how streamers can tailor their communication to maximize consumer engagement and sales outcomes.

2.2. Stereotype Theory

Stereotype theory provides an important theoretical framework for understanding how individuals form evaluations of others [35,36]. The theory suggests that people tend to judge others based on two main dimensions: warmth and competence [37]. The warmth dimension refers to individuals’ primary concern about whether others—either personally or as a group—have favorable intentions toward them. It reflects traits such as friendliness, sincerity, and trustworthiness, and indicates one’s willingness to cooperate with others. In contrast, the competence dimension addresses whether the other party has the ability to enact those intentions. It reflects characteristics such as intelligence, professionalism, and efficiency, and represents one’s perceived capability to achieve goals [19,38]. Individuals typically categorize those with a cooperative intent as high warmth, while those with a competitive intent are categorized as low warmth. Similarly, those who are able to achieve their goals are classified as high competence, while those who cannot achieve their goals are classified as low competence. In addition, stereotype theory emphasizes that there is typically a compensatory effect between warmth and competence. That is, when a social group is perceived as high on one dimension, it is often perceived as low on the other dimension [39]. For example, older adults are often perceived as high in warmth but low in competence, while wealthy individuals are typically seen as high in competence but low in warmth [38]. This compensatory perception suggests that when individuals form overall impressions of others, they weigh both warmth and competence perceptions to make a holistic judgment of their trustworthiness and intent to cooperate [40].
In commercial communication, consumers’ impressions of brands, companies, or individuals largely depend on their language expression style [41]. Language style is considered an important cue that influences how information receivers perceive others. It not only affects the emotional perception of the communicator but also shapes judgments of their competence and credibility [42]. For example, a humorous communication style is often associated with high warmth, while expression certainty is commonly linked to high competence [43]. Existing research has found that in the financial field, communicators with more certain language expression are more likely to be perceived as professional and credible [23]. In social media contexts, brands interacting with consumers using high expression certainty language can also enhance consumers’ perceptions of the brand’s competence and increase their willingness to engage [13]. In live-stream e-commerce, consumers’ impressions of the streamer largely determine their acceptance of the content recommended by the streamer and whether a purchase behavior occurs [41]. The streamer’s language style is an important source of information for consumers to assess the streamer’s warmth and competence [42], thereby influencing their trust and purchase intentions. streamers aim to present themselves as both competent and warm. Previous studies have shown that expression certainty is closely associated with the communicator’s perceived confidence and credibility [18]. Building on this foundation, the present study adopts the Stereotype Content Model as its theoretical framework to examine how streamers’ expression certainty influences consumers’ perceptions of their warmth and competence, ultimately shaping consumers’ product-related purchase decisions. The Stereotype Content Model posits that individuals tend to form impressions of others based on limited social cues, particularly in terms of perceived intent (warmth) and capability (competence). As a crucial interpersonal cue, linguistic style can effectively trigger these evaluations. Therefore, this theoretical perspective offers strong explanatory power and applicability within the context of this research.

2.3. The Impact of Streamer Expression Certainty on Consumers’ Purchase Behavior

A language style of expression certainty refers to when the communicator conveys their high confidence and certainty about the information through clear and assertive language [44,45]. In previous research on advertising language, although elements of expression certainty are often involved, most studies have focused on other linguistic style features, such as advertising tone [46], while systematic investigations of expression certainty itself remain scarce. In terms of linguistic style, advertising tone primarily emphasizes user benefits, product functionality, and the promotion of consumer action [47]. In contrast, expression certainty reflects a communicator’s strong and explicit stance toward the conveyed information. Unlike language styles that merely highlight product advantages, expression certainty allows for the presentation of both favorable and unfavorable information, thereby enhancing the objectivity and credibility of the content [13]. In live-stream e-commerce, streamers typically use a language style of expression certainty to recommend products, such as expressing their high confidence in the product through a firm tone and clear vocabulary, in order to enhance consumer trust and purchase intention [13]. Research shows that when communicators express certainty, consumers tend to perceive their information as more reliable [45], which helps reduce uncertainty in the purchasing decision and encourages consumers to make a positive purchase choice [48]. Therefore, moderate expression certainty can effectively enhance consumers’ trust in both the product and the streamer, thereby promoting purchase behavior.
However, excessive expression certainty may have the opposite effect. When communicators emphasize certainty too much, consumers may perceive it as an over-persuasion strategy, leading to doubts about the authenticity of the information and the communicator’s intentions [49]. This skeptical mindset is particularly prominent in advertising contexts, where consumers tend to resist language that overly emphasizes certainty or absolutes, perceiving it as overly focused on the communicator’s own benefit and lacking authenticity [50]. In addition, the extreme use of expression certainty may increase consumers’ cognitive load, leading to a decrease in their acceptance of the information [28]. Overall, the impact of expression certainty on consumer purchase behavior may not be a simple linear positive effect, but rather a non-linear inverted U-shaped relationship.
Based on the above analysis, this study proposes the following hypothesis:
H1: 
The effect of the streamer’s expression certainty on consumer purchase behavior follows an inverted U-shaped relationship, meaning that moderate expression certainty positively influences purchase behavior, while excessive expression certainty weakens purchase behavior.

2.4. Perceived Competence and Perceived Warmth as Mediators

According to the Stereotype Content Model, consumers typically form overall impressions of streamers based on two core dimensions: competence and warmth [37]. Specifically, the competence dimension reflects whether the streamer possesses the professionalism and intelligence required to achieve goals, while the warmth dimension captures whether the streamer is friendly, sincere, and attentive to consumers’ needs [38]. Existing research has shown that these two dimensions not only influence consumers’ attitudinal evaluations of the streamer but also directly affect their trust in the recommended information and subsequent purchase decisions [51].
When streamers use expression certainty at a moderate level, consumers tend to interpret it as a signal of competence, perceiving the streamer as knowledgeable and professional, which in turn enhances trust and stimulates purchase intentions [13]. However, the Stereotype Content Model further suggests that there is a compensation effect between competence and warmth; when an individual is perceived as more competent, their perceived warmth may relatively decline [39]. Therefore, excessively high expression certainty may reduce consumers’ perceptions of the streamer’s warmth, leading them to believe that the streamer is overly self-interested and lacks an understanding of consumer needs [50]. To further explain this mechanism, we introduce the Persuasion Knowledge Model (PKM) [52]. The model suggests that when a message is expressed with a high level of certainty, consumers are likely to perceive it as an intentional act of persuasion, thereby activating their persuasion knowledge and raising doubts about the sincerity of the message and the motives of the source [52]. When the streamer’s expression certainty is at a low or moderate level, consumers typically do not activate their persuasion defense mechanisms. In this case, an increase in expression certainty can enhance perceptions of the streamer’s competence without significantly affecting perceptions of warmth. However, when expression certainty becomes excessively high, consumers’ persuasion knowledge is activated, leading them to question the streamer’s sincerity and enthusiasm. As a result, perceived warmth significantly declines, while perceived competence does not increase further [49]. Therefore, when expression certainty is excessively high, consumers may develop a negative impression of the streamer, decreasing trust and weakening purchase behavior. Based on this, we propose the following hypotheses:
H2a: 
Perceived competence mediates the relationship between expression certainty and consumer purchase behavior when expression certainty is at a low to moderate level. Moderate increases in certainty enhance perceived competence, thereby promoting purchase behavior.
H2b: 
Perceived warmth mediates the relationship between expression certainty and consumer purchase behavior when expression certainty is at a high level. Excessively high certainty reduces perceived warmth, which in turn weakens purchase behavior.
We conducted two studies to test these hypotheses. Study 1 uses real-world data to test Hypothesis 1. Study 2 employs an experimental approach to establish causal relationships and test the potential mechanisms proposed in Hypothesis 2. Figure 1 illustrates our research design and its relationship to the hypotheses.

3. Research Method

3.1. Study 1

Study 1 analyzes real-world live-stream product consumption data and streamer language text data from the Chinese Douyin platform to test Hypothesis 1, which posits that the relationship between the streamer’s expression certainty and consumer purchase behavior follows an inverted U-shape. Douyin, one of China’s largest short-video and live-stream e-commerce platforms, currently has over 700 million daily active users and is the digital media platform with the highest user engagement time [38]. This platform not only attracts a massive consumer traffic but has also become the preferred channel for merchants and streamers to engage in product promotion and marketing collaborations, making it highly practical and influential. Therefore, analyzing live-stream data from Douyin not only effectively reflects the typical characteristics of live-stream e-commerce but also provides a representative research foundation for exploring the impact of streamers’ language expression styles on consumer behavior.

3.1.1. Data Collection

This study, based on the Douyin platform, collected live-stream video data for a total of 3842 products promoted by 143 streamers between May and July 2024, covering major consumer categories such as food, clothing, and cosmetics. The data were obtained from Huitun (www.huitun.com, accessed on 9 June 2024), a publicly accessible analytics platform that provides detailed streamer profile metrics and livestream-level user engagement statistics. A custom Python 3.10-based web crawler was developed and employed by the research team to extract multidimensional data for each product-related livestream video, including the duration of product introductions, the textual content of product recommendations, and the corresponding product sales generated directly from the video. In addition, the dataset includes basic information about each live-stream session and the corresponding products, such as the streamer’s gender, follower count, average monthly live-stream frequency, product price, and product category. Detailed sample descriptions are provided in Table 1.

3.1.2. Variables

Words are the most basic units that reflect language style. Previous studies have shown that the language expression certainty can be measured by calculating the proportion of certainty-related words in the text [40]. Referencing the study by Pezzuti et al. [13], this paper uses the widely used language query and word count software (Linguistic Inquiry and Word Count, LIWC 22) in the field of text analysis to quantify the expression certainty in the language content of each product’s live-stream by the streamer. The LIWC software includes 80 built-in dictionaries, which use word substitution methods to identify specific vocabulary in texts and calculate the proportion of these words within the overall text, thereby extracting textual features [40]. The certainty dictionary contains 113 words that describe certainty, such as “absolute”, “must”, “never”, “perfect”, etc. [41]. This study utilizes the Chinese version of the LIWC dictionary developed by Huang et al. [38], which was adapted from the LIWC 2007 framework to suit the linguistic and cultural characteristics of Chinese texts. This Chinese-language LIWC dictionary has been widely applied in various studies involving Chinese text analysis and has demonstrated good applicability and validity [38,39]. For example, previous studies employed this Chinese version of the LIWC dictionary to extract linguistic features in their studies, confirming its stability and reliability in Chinese language contexts [38,39]. Therefore, using the Chinese version of LIWC to extract expression certainty features in this study is theoretically well-founded and empirically supported.
In this study, the live-stream data sample showed that 98.5% of the product live streams featured at least one certainty-related word. This indicates that the expression certainty is highly prevalent in live-stream language. Ultimately, this study uses LIWC to calculate the proportion of certainty-related words in the streamer’s language content during each product’s live stream, serving as an indicator of the streamer’s expression certainty. This method not only provides a standardized measurement tool for text analysis but also enhances the objectivity and reliability of the research results.
This study uses the product sales during each live stream as the dependent variable—consumer purchase behavior [38]. In this way, the study accurately measures the impact of the streamer’s expression certainty on purchasing behavior in real-world scenarios.
To rule out potential confounding effects, this study controlled for several influencing factors based on prior literature. First, Pezzuti et al. suggest that a streamer’s gender may affect the persuasiveness of different information strategies [13]. Accordingly, gender was included as a dummy variable (1 = male; 0 = female). Second, the number of followers serves as a proxy for a streamer’s popularity and significantly affects consumers’ perceptions of their competence [38]. To account for this, follower count was operationalized as a set of binary-coded dummy variables: (0,0) for fewer than 100,000 followers, (0,1) for 100,000–1 million, (1,0) for 1–10 million, and (1,1) for more than 10 million. In addition, a streamer’s frequency of live stream reflects their overall activity level and interaction frequency with consumers. Following Wongkitrungrueng et al. [53], we controlled for this factor using the average number of live-stream sessions per month. Product price was also controlled for, as pricing may directly affect sales performance. Similarly, we accounted for the duration allocated to each product during the livestream, as longer presentation time may influence consumers’ attention allocation and depth of information processing. Regarding message characteristics, we controlled for the readability of the streamer’s language content. Readability was assessed using the sentence generation probability method proposed by Shin et al., which employs a Word2vec-based deep learning algorithm. This approach considers the linguistic and semantic context of vocabulary, calculating sentence-level generation probabilities based on the frequency of infrequent words in a pretrained corpus to reflect textual complexity. Furthermore, to account for the potential influence of information load, we controlled for the total word count of each product-related language segment, following Yang et al. [32]. We also controlled for the frequency of money-related language used by streamers. This variable captures the extent to which streamers rely on pricing or incentive-based strategies in their messaging, which may affect consumer purchase decisions [24]. Lastly, we accounted for social-related language, which refers to the degree to which streamers adopt social interaction strategies during livestreams. This was measured by calculating the frequency of socially oriented words, following the method outlined in Tausczik and Pennebaker [24], as such strategies have been shown to influence consumer behavior.

3.1.3. Data Analysis

To ensure the validity of the analysis of secondary data in this study, we first evaluated the statistical distribution of the dataset to select an appropriate econometric model [24]. The statistical results show that in this dataset, product sales are a non-negative integer set, making it count data that follows a positively skewed distribution. The increase in product sales for each product live-stream video ranges from 0 to 166,524 (M = 652, SD = 6268.28; Skewness = 28.42, SE = 0.082). For count-type data that is not normally distributed, previous research has shown that traditional ordinary least squares (OLS) regression performs poorly [42,43]. Therefore, following recommendations from previous studies, this research uses negative binomial regression to analyze the dataset, as this method has been proven to effectively handle overdispersed data [24].
To verify the inverted U-shape relationship, we used quadratic regression, as it is a foundational method for testing U-shaped relationships [54]. The formula for the quadratic regression is as follows:
y = m + a x + b x 2 + ε   C o n t r o l s
Here, x is the independent variable, x 2 is the quadratic term, m is the intercept, C o n t r o l s is the vector of all control variables, and ε represents the coefficients of the control variables. If the coefficient b of the quadratic term is negative and significant, it indicates an inverted U-shape relationship. Additionally, some researchers suggest further validating the inverted U-shape relationship by checking whether the turning point x m a x = − a /2 b lies within the possible range of x values [55]. Therefore, in verifying the inverted U-shape relationship, we checked whether the turning point lies within a reasonable range.

3.1.4. Results

This study fitted three models using negative binomial regression to test the impact of expression certainty on consumer purchase behavior, as shown in Table 2. To exclude the influence of multicollinearity on the research conclusions, the variance inflation factor (VIF) for each variable was calculated. The results show that the maximum VIF value of the variables is 3.95, which is below the threshold of 5, indicating that there is no severe multicollinearity issue in this study [13].
First, Model 1 examines the relationship between all control variables and the dependent variable. Model 2 adds expression certainty to Model 1, aiming to test the main effect of expression certainty on purchase behavior. The results show that the likelihood ratio test for the negative binomial regression is significant (χ2(10) = −6442.37, p < 0.001), indicating that the model has a good fit. Moreover, expression certainty has a significant positive effect on purchase behavior (β = 0.241, p < 0.001), which means that the streamer’s expression certainty can promote consumer purchase behavior. Next, Model 3 tests whether there is an inverted U-shaped relationship between expression certainty and purchase behavior through quadratic regression. Based on Model 2, the squared term of expression certainty was added. The likelihood ratio test for Model 3 is significant (χ2(11) = −6328.24, p < 0.001), and the AIC value significantly decreased, indicating an improved model fit. At the same time, the squared term of expression certainty has a significantly negative coefficient (β = −0.143, p < 0.001), indicating that there is an inverted U-shaped relationship between the streamer’s expression certainty and consumer purchase behavior. Subsequently, we further examined the turning point to support the inverted U-shaped relationship. As shown in Figure 2, the turning point for expression certainty is 0.84, which lies in the middle of the observed values, further supporting the inverted U-shaped influence of the streamer’s expression certainty on consumer purchase behavior.

3.1.5. Robustness Test

In the robustness test, we conducted two stages of analysis. First, we tested the stability of the dataset. Due to differences in the measurement units of different explanatory variables, we standardized the explanatory variables so that the mean of each variable was 0 and the standard deviation was 1. Based on this analysis, we decided to use Ordinary Least Squares (OLS) to examine the inverted U-shaped effect of the streamer’s expression certainty on consumer purchase behavior.
Second, we conducted a stability test on the dependent variable. In the dataset, some products had zero sales during the live stream. To eliminate the impact of these zero-purchase data on the results, we excluded 295 samples and tested the remaining data. The results, shown in Table 3, indicate that the results of the two robustness test again confirm our hypothesis. These robustness test results significantly enhance the reliability of the study conclusions.

3.2. Study 2

Study 2 aims to extend the findings of Study 1 in two meaningful ways. First, Study 1 provided external validity for the curvilinear effect of the streamer’s expression certainty on consumer behavior using real-world live streaming data. However, due to the nature of the data, we were unable to establish causal relationships. Study 2 addresses this issue through an experimental method. Second, we proposed that the curvilinear effect is driven by perceived competence and warmth (H2a and H2b). Study 2 will explicitly test this hypothesis.

3.2.1. Method

Study 2 adopts a between-subject experimental design with a single factor (expression certainty: low, moderate, high), aimed at testing Hypotheses H1 and H2a and H2b—namely, that the streamer’s expression certainty has an inverted U-shaped effect on consumer purchase behavior, and that perceived competence and perceived warmth mediate this effect. Study 2 recruited a total of 240 participants (Mage = 31.61; 63.75% female) through the online data collection platform Credamo (www.credamo.com, accessed on 1 July 2024) and employed a scenario-based experiment to collect the data. Participants were asked to evaluate a live-stream script, which was adapted and shortened from a real live stream. The script features a virtual skincare product to ensure participants can generally understand and evaluate the product. The live-stream script shown to participants in each group differed only in the level of expression certainty, while all other information remained consistent. We manipulated expression certainty by changing certain words. In the low expression certainty group, participants read “This skincare product works for some people. It contains important ingredients like hyaluronic acid and Vitamin C, which may help with skin issues. Maybe you could try it, the results should be good”. In the medium expression certainty group, participants read “This skincare product works for everyone. It contains important ingredients like hyaluronic acid and Vitamin C, which may help with skin issues. Maybe you could try it, the results will definitely be good”. In the high expression certainty group, participants read “This skincare product works for everyone. It contains important ingredients like hyaluronic acid and Vitamin C, which will definitely improve skin issues. You must try it, the results will definitely be good”. In all three conditions, the number of certain and tentative words was the same (i.e., four each). We manipulated the relative proportion of certain and tentative words under each condition: four tentative words in the low condition, two certain words and two tentative words in the medium condition, and four certain words in the high condition. This study recruited 90 participants (57.7% female; Mage = 31.25) to take part in a pretest aimed at assessing the effectiveness of our manipulation of expression certainty in the video stimuli. After viewing the experimental materials, participants rated the level of expression certainty using a single-item 7-point scale (1 = completely uncertain, 7 = completely certain). A one-way ANOVA revealed significant differences in perceived expression certainty across the three conditions (M high = 6.24, SD = 1.18; M medium = 5.38, SD = 1.12; M low = 4.21, SD = 1.24; F(2, 87) = 38.52, p < 0.001), indicating that the manipulation was successful and the materials were suitable for use in the main experiment.
In the formal experiment, participants were randomly assigned to one of three groups based on different levels of expression certainty. After reading the live-stream script, they were asked to rate their purchase intention on a 7-point scale (1 = not willing to buy at all; 7 = very willing to buy). They also rated the streamer’s perceived competence (“competent”, “capable”, and “skilled”; Cronbach’s α = 0.93) and warmth (“enthusiastic”, “sincere”, and “friendly”; Cronbach’s α = 0.94) on a 7-point scale (1 = not at all, 7 = very much; adapted from Chang et al. [56]; Fiske et al. [19]). Participants also rated the streamer’s expression certainty on a 7-point scale (1 = completely uncertain, 7 = completely certain). At the end of the survey, participants were asked to provide their age and gender.

3.2.2. Results Analysis and Discussion

1. Manipulation Check
First, a manipulation check was conducted for the streamer’s expression certainty in Study 2. One-way analysis of variance (ANOVA) revealed a significant effect (F(2, 237) = 14.35, p < 0.001). Participants in the high-certainty group (M = 5.37, SD = 1.33) rated the streamer’s expression certainty higher than those in the moderate- (M = 4.46, SD = 1.54; Fisher’s LSD, p = 0.007) and low-certainty groups (M = 3.58, SD = 1.27; Fisher’s LSD, p < 0.001). Additionally, participants in the moderate-certainty group rated the streamer’s expression certainty higher than those in the low-certainty group (Fisher’s LSD, p = 0.007). These results indicate that the manipulation of expression certainty in Study 2 was successful.
2. Hypothesis Testing
First, the main effect of expression certainty on purchase intention was analyzed. One-way ANOVA showed a significant effect (F(2, 237) = 4.29, p = 0.015). Participants in the moderate-certainty group (M = 5.25, SD = 1.48) reported higher purchase intention than those in the high-certainty group (M = 4.48, SD = 1.44; Fisher’s LSD, p = 0.017) and the low-certainty group (M = 4.25, SD = 1.38; Fisher’s LSD, p = 0.022). There was no significant difference in purchase intention scores between the high- and low-certainty groups (Fisher’s LSD, p = 0.725). These results further support Hypothesis 1.
Next, we performed a one-way ANOVA to test the effects of expression certainty on the streamer’s perceived competence (F(2, 237) = 7.52, p < 0.001) and warmth (F(2, 237) = 23.22, p < 0.001). Consistent with our expectations, participants in the moderate (M = 5.67, SD = 1.36; Fisher’s LSD, p = 0.013) and high-certainty groups (M = 5.72, SD = 1.44; Fisher’s LSD, p = 0.018) rated the streamer’s perceived competence higher than participants in the low-certainty group (M = 3.75, SD = 1.24). There was no significant difference in perceived competence between the moderate and high-certainty groups (Fisher’s LSD, p = 0.529).
In contrast, participants in the moderate-certainty group (M = 4.88, SD = 1.42; Fisher’s LSD, p < 0.001) and low-certainty group (M = 4.74, SD = 1.25; Fisher’s LSD, p < 0.001) rated the streamer’s perceived warmth higher than participants in the high-certainty group (M = 3.42, SD = 1.21). There was no significant difference in perceived warmth between the moderate- and low-certainty groups (Fisher’s LSD, p = 0.277) (see Figure 3 for detailed results).
For testing the mediating effects of perceived competence and perceived warmth, we conducted a mediation analysis using PROCESS Model 4 with 5000 bootstrap resamples [13]. By comparing the data of the moderate-level group with the low-level group, we found that the indirect effect of expression certainty through perceived competence was significant (B = 0.21, SE = 0.09, 95% CI: 0.07–0.37), while the indirect effect through perceived warmth was not significant (B = −0.06, SE = 0.08, 95% CI: −0.18–0.12). Conversely, by comparing the data of the moderate-level group with the high-level group, we found that the indirect effect of expression certainty through perceived warmth was significant (B = 0.31, SE = 0.07, 95% CI: 0.24–0.52), while the indirect effect through perceived competence was not significant (B = 0.04, SE = 0.08, 95% CI: −0.14–0.16). These findings provide evidence for Hypothesis 2, suggesting that the curvilinear relationship between the streamer’s expression certainty and consumer purchase behavior can be explained by a trade-off between perceived competence and warmth. Specifically, our results show that when the streamer’s expression certainty increases from low to moderate levels, the perceived competence of the streamer increases in the eyes of the consumer, but perceived warmth does not decrease. However, when the streamer’s expression certainty increases from moderate to high levels, consumers’ perceived warmth decreases, but perceived competence does not increase further. Given that both perceived competence and warmth can predict consumer purchase behavior, moderate expression certainty is the most ideal for promoting purchase behavior. See Figure 4 for detailed results.

4. Discussion

This study empirically analyzes 3842 product live-stream data from the Douyin platform and conducts an experimental study to explore the impact of streamers’ expression certainty on consumer purchasing behavior and its underlying mechanisms. The results show that streamers’ expression certainty has a significant inverted U-shaped effect on consumer purchasing behavior, with perceived competence and perceived warmth serving as mediators in this process. These findings offer important insights for both theoretical research and practical applications.
Firstly, this study finds that streamers’ expression certainty has an inverted U-shaped relationship with consumer purchasing behavior. Moderate levels of expression certainty enhance consumer trust and purchase intention, while excessively high or low expression certainty weakens this effect. This result aligns with existing literature that suggests expression certainty conveys confidence and enhances persuasiveness [57]. For instance, previous studies have shown that language with high expression certainty increases consumer engagement with brand information on social media [13]. The findings of this study further indicate that, in the context of live-stream e-commerce, more expression certainty is not always better; excessive certainty may trigger consumer skepticism, thereby suppressing purchase behavior. The results support Hypothesis 1.
Furthermore, this study verifies the mediating role of perceived competence and perceived warmth between expression certainty and consumer purchasing behavior. Consumers typically assess a streamer’s competence through language cues [58], and a language style with high expression certainty is associated with the stereotype of high competence, leading consumers to perceive streamers using such language as more confident and professional [48]. In addition, perceived warmth is an important factor influencing consumer trust. Moderate expression certainty helps maintain the streamer’s image of sincerity and friendliness in the eyes of consumers, thereby enhancing their purchase intention. This finding aligns with the compensatory effect of stereotype theory [22], further suggesting that streamers need to strike a balance between expressing competence and warmth. The results support Hypotheses 2a and 2b. In addition, regarding the compensation effect within the Stereotype Content Model, this study further reveals that the compensatory relationship between perceived warmth and perceived competence exhibits heterogeneity across different competence levels. When consumers perceive the streamer’s competence to be relatively low, the compensation effect between competence and warmth is not yet prominent; that is, an increase in perceived competence does not lead to a significant decrease in perceived warmth. However, when perceived competence reaches a higher level, the compensation effect becomes more pronounced. At this stage, consumers are more likely to cognitively trade off competence and warmth, resulting in a significant decline in warmth perceptions. This finding enriches the understanding of the dynamic changes in the compensation effect under different competence perception scenarios.
Finally, this study emphasizes the distinct roles of perceived competence and perceived warmth in this process. While high expression certainty enhances consumers’ perception of the streamer’s competence, excessively high expression certainty may weaken the perception of the streamer’s warmth. This negative effect is particularly evident in consumers’ skepticism about the streamer’s true intentions, further affecting their trust and behavioral responses.

4.1. Theoretical Implications

Firstly, this study enriches the research field of streamer language style. By exploring the effect of streamer expression certainty on consumer purchasing behavior, this study reveals the critical role of language expression style in consumer behavior in live-stream e-commerce. Previous research has mostly focused on streamer source traits (such as credibility and attractiveness) and their impact on consumer behavior [59,60]. However, the language characteristics of streamers are key to the success of live-stream marketing [61]. This study adopts a novel and practically relevant perspective by focusing on a specific linguistic feature—expression certainty—and investigates how it influences consumer purchase behavior and intention. In doing so, it extends existing research on the impact of streamer language styles on consumer responses.
Secondly, this study provides new empirical support for stereotype theory. Previous literature has mainly focused on how consumers’ stereotypes about competence and warmth affect their social judgments [36], with less attention given to the influence of language style on this process. Although existing research suggests that expression certainty is associated with confidence and high competence [38], marketing scholars have yet to explore the relationship between expression certainty and purchasing behavior, along with the underlying mechanisms. This study uncovers the link between expression certainty and perceived competence and warmth, offering a new theoretical perspective on understanding the relationship between streamer language style and stereotypes. It further demonstrates that streamer language style can influence consumer purchasing behavior by affecting perceived competence and warmth.
Finally, this study extends the understanding of consumer behavior in the live-stream e-commerce domain. By emphasizing the mediating roles of perceived competence and perceived warmth, this study reveals the deeper impact mechanisms of streamer language expression strategies on consumer behavior. Specifically, this study verifies the curvilinear relationship between expression certainty and consumer behavior, highlighting the need for streamers to balance the appropriateness of language expression when enhancing consumer trust and purchase intention. This finding provides a theoretical foundation for optimizing streamer language strategies in live-stream e-commerce practice.

4.2. Practical Implications

This study provides the following important practical implications for streamers and businesses in live-stream e-commerce.
Firstly, the results of this study show that the effect of streamer expression certainty on consumer purchasing behavior follows an inverted U-shaped curve. This finding reminds businesses and streamers that the strategic use of language expression is a key way to optimize live-stream outcomes. In practice, streamers should avoid both excessively low and high levels of expression certainty. For instance, low expression certainty may weaken consumer trust in the information due to its ambiguity, while excessive expression certainty may raise doubts about the streamer’s true intentions. Therefore, when recommending products, streamers should maintain expression certainty within a moderate range, using clear, confident, but not excessive language to convey professionalism and credibility, thereby maximizing consumer purchasing behavior.
Secondly, this study reveals the important mediating role of perceived competence and perceived warmth in the relationship between expression certainty and consumer purchasing behavior, which has significant implications for optimizing streamer language style. Streamers should not only demonstrate professional competence in their language expression but also maintain consumers’ perceptions of their sincerity and friendliness. For example, streamers can express confidence moderately, address practical issues that consumers care about, and avoid overly absolute language, thus balancing perceptions of competence and warmth. Such language strategies can enhance consumer trust and likability, further improving the effectiveness of live-streamed recommendations.
Finally, this study provides a reference direction for businesses in designing streamer training programs. Businesses can help streamers master moderate expression certainty through language strategy training, while also enhancing their professional knowledge and communication skills to improve consumers’ overall perception of their competence and warmth. Depending on the type of product, businesses can also guide streamers to adjust their language style based on the preferences of their target audience, achieving better live-stream promotion outcomes.

4.3. Research Limitations and Future Research Avenues

This study has several limitations that warrant further exploration in future research. First, this study primarily analyzes live-stream data from the Douyin platform. Although Douyin, as one of the largest live-stream e-commerce platforms in China, is highly representative, the user demographics and content consumption patterns may differ across other social media platforms (e.g., YouTube or Twitch). Future research could apply the model developed in this study to other platforms to verify the generalizability of the impact of expression certainty on consumer behavior. Second, this study uses a predefined dictionary (LIWC) to measure the level of certainty in the streamer’s language expression. While this method is widely used in text analysis, it does not distinguish between the specific impact of different certainty-related words on consumer behavior. In addition, considering the complexity of the Chinese language, although a localized version of the LIWC dictionary has been developed [38], it has not been updated as promptly or comprehensively as the English version. Therefore, future research could incorporate more flexible Natural Language Processing (NLP) approaches, such as machine learning or deep learning techniques, to more accurately identify different types of certainty expressions in the Chinese context and further explore the heterogeneous effects of various certainty-related terms on consumer behavior.
Third, this study mainly focuses on the impact of expression certainty on consumer purchasing behavior and does not address other important consumer behaviors or psychological variables. For example, expression certainty may influence consumers’ loyalty to streamers, their attitudes toward product recommendations, or their overall brand perception—variables that are also important in real-world marketing decisions. Future research could expand the range of dependent variables, exploring the multi-layered effects of streamer language style from a more comprehensive perspective, providing richer theoretical guidance and practical support for businesses and streamers.

Author Contributions

Conceptualization, Y.C. and Q.H.; methodology, Y.J. and Q.H.; software, Z.L.; validation, Y.C., Q.H. and Y.J.; formal analysis, Y.C.; investigation, Z.L.; resources, W.Z.; data curation, W.Z.; writing—original draft preparation, Y.C. and Q.H.; writing—review and editing, Y.C. and Q.H.; visualization, Z.L. and Y.J.; supervision, W.Z. and Y.J.; project administration, Y.J.; funding acquisition, Q.H. All authors have read and agreed to the published version of the manuscript.

Funding

This research was funded by the National Natural Science Foundation of China—Young Scientists Fund Project, grant number 72402097, Fujian Provincial Social Science Foundation, grant number FJ2022BF036, Postgraduate Research & Practice Innovation Program of Jiangsu Province, grant number KYCX25_2466, and Postgraduate Research & Practice Innovation Program of Jiangsu Province, grant number SJCX25_1132. And The APC was funded by authors.

Institutional Review Board Statement

The study was conducted in accordance with the Declaration of Helsinki, and approved by the Academic Ethics Committee of Nanjing Audit University (protocol code 2024NAUEC008, date of approval 15 March 2024).

Informed Consent Statement

Informed consent was obtained from all subjects involved in the study.

Data Availability Statement

The original contributions presented in the study are included in the article, further inquiries can be directed to the corresponding author/s.

Conflicts of Interest

The authors declare no conflict of interest.

References

  1. Sun, J.; Yang, J.; Wang, Y. Vertically versus horizontally differentiated information disclosure in travel live streams–the role of sensory imagery. J. Res. Interact. Mark. 2022, 17, 353–373. [Google Scholar] [CrossRef]
  2. Wang, C.L. Editorial–What is an interactive marketing perspective and what are emerging research areas? J. Res. Interact. Mark. 2024, 18, 161–165. [Google Scholar] [CrossRef]
  3. Gu, X.; Zhang, X.; Kannan, P. Influencer mix strategies in livestream commerce: Impact on product sales. J. Mark. 2024, 88, 64–83. [Google Scholar] [CrossRef]
  4. Wu, R.; Liu, J.; Chen, S.; Tong, X. The effect of E-commerce virtual live streamer socialness on consumers’ experiential value: An empirical study based on Chinese E-commerce live streaming studios. J. Res. Interact. Mark. 2023, 17, 714–733. [Google Scholar] [CrossRef]
  5. Chen, N.; Yang, Y. The role of influencers in live streaming e-commerce: Influencer trust, attachment, and consumer purchase intention. J. Theor. Appl. Electron. Commer. Res. 2023, 18, 1601–1618. [Google Scholar] [CrossRef]
  6. Bai, S.; Jiang, F.; Li, Q.; Yu, D.; Tan, Y. Harnessing Empathy: The Power of Emotional Resonance in Live Streaming Sales and the Moderating Magic of Product Type. J. Theor. Appl. Electron. Commer. Res. 2025, 20, 30. [Google Scholar] [CrossRef]
  7. Wan, Q.; Chen, J.; Yu, C.; Lu, M.; Liu, D. Optimal marketing strategies for live streaming rooms in livestream e-commerce. Electron. Commer. Res. 2024, 1–34. [Google Scholar] [CrossRef]
  8. Kim, J.; He, N.; Miles, I. Live Commerce Platforms: A New Paradigm for E-Commerce Platform Economy. J. Theor. Appl. Electron. Commer. Res. 2023, 18, 959–975. [Google Scholar] [CrossRef]
  9. Yao, X.; Ren, W. Enhancing Customer Experience: A Holistic Brand Strategy in the Metaverse. J. Appl. Bus. Behav. Sci. 2025, 1, 115–135. [Google Scholar] [CrossRef]
  10. Cascio Rizzo, G.L.; Villarroel Ordenes, F.; Pozharliev, R.; De Angelis, M.; Costabile, M. How high-arousal language shapes micro-versus macro-influencers’ impact. J. Mark. 2024, 88, 00222429231207636. [Google Scholar] [CrossRef]
  11. Liao, J.; Chen, K.; Qi, J.; Li, J.; Yu, I.Y. Creating immersive and parasocial live shopping experience for viewers: The role of streamers’ interactional communication style. J. Res. Interact. Mark. 2023, 17, 140–155. [Google Scholar] [CrossRef]
  12. Jiao, Y.; Sarigöllü, E.; Lou, L.; Huang, B. How streamers foster consumer stickiness in live streaming sales. J. Theor. Appl. Electron. Commer. Res. 2023, 18, 1196–1216. [Google Scholar] [CrossRef]
  13. Pezzuti, T.; Leonhardt, J.M.; Warren, C. Certainty in language increases consumer engagement on social media. J. Interact. Mark. 2021, 53, 32–46. [Google Scholar] [CrossRef]
  14. Wang, C.L. Editorial–The misassumptions about contributions. J. Res. Interact. Mark. 2022, 16, 1–2. [Google Scholar] [CrossRef]
  15. Jiang, Y.; Lei, J. I Will Recommend A Salad but Choose A Burger for You: The Effect of Decision Tasks and Social Distance on Food Decisions for Others. J. Appl. Bus. Behav. Sci. 2025, 1, 95–114. [Google Scholar] [CrossRef]
  16. Wang, C.L. Demonstrating contributions through storytelling. J. Res. Interact. Mark. 2025, 19, 1–4. [Google Scholar]
  17. Karmarkar, U.R.; Tormala, Z.L. Believe me, I have no idea what I’m talking about: The effects of source certainty on consumer involvement and persuasion. J. Cons. Res. 2010, 36, 1033–1049. [Google Scholar] [CrossRef]
  18. Moy, N.; Chan, H.F.; Septianto, F.; Mathmann, F.; Torgler, B. Confidence is Good? too Much, not so Much: Exploring the effects on crowdfunding success. J. Bus. Res. 2024, 182, 114711. [Google Scholar] [CrossRef]
  19. Fiske, S.T. Stereotype content: Warmth and competence endure. Curr. Dir. Psychol. Sci. 2018, 27, 67–73. [Google Scholar] [CrossRef]
  20. Davvetas, V.; Halkias, G. Global and local brand stereotypes: Formation, content transfer, and impact. Int. Mark. Rev. 2019, 36, 675–701. [Google Scholar] [CrossRef]
  21. Judd, C.M.; James-Hawkins, L.; Yzerbyt, V.; Kashima, Y. Fundamental dimensions of social judgment: Understanding the relations between judgments of competence and warmth. J. Pers. Soc. Psychol. 2005, 89, 899. [Google Scholar] [CrossRef] [PubMed]
  22. Yzerbyt, V.Y.; Kervyn, N.; Judd, C.M. Compensation versus halo: The unique relations between the fundamental dimensions of social judgment. Pers. Soc. Psychol. Bull. 2008, 34, 1110–1123. [Google Scholar] [CrossRef]
  23. Bryan, A.D.; Webster, G.D.; Mahaffey, A.L. The big, the rich, and the powerful: Physical, financial, and social dimensions of dominance in mating and attraction. Pers. Soc. Psychol. Bull. 2011, 37, 365–382. [Google Scholar] [CrossRef]
  24. Tausczik, Y.R.; Pennebaker, J.W. The psychological meaning of words: LIWC and computerized text analysis methods. J. Consum. Behav. 2010, 29, 24–54. [Google Scholar] [CrossRef]
  25. Keeling, K.; McGoldrick, P.; Beatty, S. Avatars as salespeople: Communication style, trust, and intentions. J. Bus. Res. 2010, 63, 793–800. [Google Scholar] [CrossRef]
  26. Roy, R.; Naidoo, V. Enhancing chatbot effectiveness: The role of anthropomorphic conversational styles and time orientation. J. Bus. Res. 2021, 126, 23–34. [Google Scholar] [CrossRef]
  27. Liu, L.; Fang, J.; Yang, L.; Han, L.; Hossin, M.A.; Wen, C. The power of talk: Exploring the effects of streamers’ linguistic styles on sales performance in B2B livestreaming commerce. Inf. Process. Manag. 2023, 60, 103259. [Google Scholar] [CrossRef]
  28. Luo, H.; Cheng, S.; Zhou, W.; Yu, S.; Lin, X. A study on the impact of linguistic persuasive styles on the sales volume of live streaming products in social e-commerce environment. Mathematics 2021, 9, 1576. [Google Scholar] [CrossRef]
  29. Xu, Y.; Zhang, J.; Deng, G. Enhancing customer satisfaction with chatbots: The influence of communication styles and consumer attachment anxiety. Front. Psychol. 2022, 13, 902782. [Google Scholar] [CrossRef]
  30. Chattaraman, V.; Kwon, W.-S.; Gilbert, J.E.; Ross, K. Should AI-Based, conversational digital assistants employ social-or task-oriented interaction style? A task-competency and reciprocity perspective for older adults. Comput. Hum. Behav. 2019, 90, 315–330. [Google Scholar] [CrossRef]
  31. Zhang, J.; Chen, Q.; Lu, J.; Wang, X.; Liu, L.; Feng, Y. Emotional expression by artificial intelligence chatbots to improve customer satisfaction: Underlying mechanism and boundary conditions. Tour. Manag. 2024, 100, 104835. [Google Scholar] [CrossRef]
  32. Yang, N.; Wang, Z. Addressing as a gender-preferential way for suggestive selling in Chinese e-commerce live streaming discourse: A corpus-based approach. J. Pragmat. 2022, 197, 43–54. [Google Scholar] [CrossRef]
  33. Liu, Z.; Lei, S.-h.; Guo, Y.-l.; Zhou, Z.-a. The interaction effect of online review language style and product type on consumers’ purchase intentions. Palgrave Commun. 2020, 6, 11. [Google Scholar] [CrossRef]
  34. Van Dolen, W.M.; Dabholkar, P.A.; De Ruyter, K. Satisfaction with online commercial group chat: The influence of perceived technology attributes, chat group characteristics, and advisor communication style. J. Retail. 2007, 83, 339–358. [Google Scholar] [CrossRef]
  35. Brough, A.R.; Wilkie, J.E.; Ma, J.; Isaac, M.S.; Gal, D. Is eco-friendly unmanly? The green-feminine stereotype and its effect on sustainable consumption. J. Cons. Res. 2016, 43, 567–582. [Google Scholar] [CrossRef]
  36. Kervyn, N.; Fiske, S.T.; Malone, C. Social perception of brands: Warmth and competence define images of both brands and social groups. Consum. Psychol. Rev. 2022, 5, 51–68. [Google Scholar] [CrossRef]
  37. Daniels, M.E.; Samper, A.; Morales, A.C. What Do You Do Outside of Work? When and Why Disclosing Eudaimonic Leisure Activities Increases Competence Perceptions and Consumer Choice. J. Cons. Res. 2025, ucaf004. [Google Scholar] [CrossRef]
  38. Yang, Q.; Huo, J.; Li, H.; Xi, Y.; Liu, Y. Can social interaction-oriented content trigger viewers’ purchasing and gift-giving behaviors? Evidence from live-streaming commerce. Internet Res. 2023, 33, 46–71. [Google Scholar] [CrossRef]
  39. Yang, Q.; Li, H.; Lin, Y.; Jiang, Y.; Huo, J. Fostering consumer engagement with marketer-generated content: The role of content-generating devices and content features. Internet Res. 2022, 32, 307–329. [Google Scholar] [CrossRef]
  40. Zeileis, A.; Kleiber, C.; Jackman, S. Regression models for count data in R. J. Stat. Softw. 2008, 27, 1–25. [Google Scholar] [CrossRef]
  41. Curran, P.J.; West, S.G.; Finch, J.F. The robustness of test statistics to nonnormality and specification error in confirmatory factor analysis. Psychol. Methods 1996, 1, 16. [Google Scholar] [CrossRef]
  42. O’brien, R.M. A caution regarding rules of thumb for variance inflation factors. Qual. Quant. 2007, 41, 673–690. [Google Scholar] [CrossRef]
  43. Artigas, E.M.; Vilches-Montero, S.; Yrigoyen, C.C. Antecedents of tourism destination reputation: The mediating role of familiarity. J. Retail. Consum. Serv. 2015, 26, 147–152. [Google Scholar] [CrossRef]
  44. Corley, P.C.; Wedeking, J. The (dis) advantage of certainty: The importance of certainty in language. Law Soc. Rev. 2014, 48, 35–62. [Google Scholar] [CrossRef]
  45. Han, S.-H.; Lind, C.J. Putting powerfulness in its place: A study on discursive style in public discussion and its impact. Argum. Adv. 2017, 53, 216–233. [Google Scholar] [CrossRef]
  46. Barcelos, R.H.; Dantas, D.C.; Sénécal, S. Watch your tone: How a brand’s tone of voice on social media influences consumer responses. J. Interact. Mark. 2018, 41, 60–80. [Google Scholar] [CrossRef]
  47. Pogacar, R.; Shrum, L.; Lowrey, T.M. The effects of linguistic devices on consumer information processing and persuasion: A language complexity× processing mode framework. J. Consum. Psychol. 2018, 28, 689–711. [Google Scholar] [CrossRef]
  48. De Valck, K.; Van Bruggen, G.H.; Wierenga, B. Virtual communities: A marketing perspective. Bus. Horiz. 2009, 47, 185–203. [Google Scholar] [CrossRef]
  49. Song, S.; Xu, Y.; Ma, B.; Zong, X. Understanding How Streamer’s Self-Presentation in E-Commerce Live Streaming Affects Consumers: The Role of Persuasion Knowledge. J. Theor. Appl. Electron. Commer. Res. 2024, 19, 1922–1942. [Google Scholar] [CrossRef]
  50. Krishnan, B.C.; Dutta, S.; Jha, S. Effectiveness of exaggerated advertised reference prices: The role of decision time pressure. J. Retail. 2013, 89, 105–113. [Google Scholar] [CrossRef]
  51. Yao, R.; Qi, G.; Wu, Z.; Sun, H.; Sheng, D. Digital human calls you dear: How do customers respond to virtual streamers’ social-oriented language in e-commerce livestreaming? A stereotyping perspective. J. Retail. Consum. Serv. 2024, 79, 103872. [Google Scholar] [CrossRef]
  52. Friestad, M.; Wright, P. The persuasion knowledge model: How people cope with persuasion attempts. J. Cons. Res. 1994, 21, 1–31. [Google Scholar] [CrossRef]
  53. Wongkitrungrueng, A.; Dehouche, N.; Assarut, N. Live streaming commerce from the sellers’ perspective: Implications for online relationship marketing. J. Mark. Manag. 2020, 36, 488–518. [Google Scholar] [CrossRef]
  54. Simonsohn, U. Two lines: A valid alternative to the invalid testing of U-shaped relationships with quadratic regressions. Adv. Methods Pract. Psychol. Sci. 2018, 1, 538–555. [Google Scholar] [CrossRef]
  55. Haans, R.F.; Pieters, C.; He, Z.L. Thinking about U: Theorizing and testing U-and inverted U-shaped relationships in strategy research. Strateg. Manag. J. 2016, 37, 1177–1195. [Google Scholar] [CrossRef]
  56. Chang, Y.; Li, Y.; Yan, J.; Kumar, V. Getting more likes: The impact of narrative person and brand image on customer–brand interactions. J. Acad. Mark. Sci. 2019, 47, 1027–1045. [Google Scholar] [CrossRef]
  57. Rucker, D.D.; Tormala, Z.L.; Petty, R.E.; Briñol, P. Consumer conviction and commitment: An appraisal-based framework for attitude certainty. J. Consum. Psychol. 2014, 24, 119–136. [Google Scholar] [CrossRef]
  58. Alsaad, A.K. Ethical judgment, subjective norms, and ethical consumption: The moderating role of moral certainty. J. Retail. Consum. Serv. 2021, 59, 102380. [Google Scholar] [CrossRef]
  59. Ashraf, A.; Hameed, I.; Saeed, S.A. How do social media influencers inspire consumers’ purchase decisions? The mediating role of parasocial relationships. Int. J. Consum. Stud. 2023, 47, 1416–1433. [Google Scholar] [CrossRef]
  60. Zhang, Q.; Ariffin, S.K.; Richardson, C.; Wang, Y. Influencing factors of customer loyalty in mobile payment: A consumption value perspective and the role of alternative attractiveness. J. Retail. Consum. Serv. 2023, 73, 103302. [Google Scholar] [CrossRef]
  61. Luo, L.; Xu, M.; Zheng, Y. Informative or affective? Exploring the effects of streamers’ topic types on user engagement in live streaming commerce. J. Retail. Consum. Serv. 2024, 79, 103799. [Google Scholar] [CrossRef]
Jtaer 20 00138 g001
Figure 1. Conceptual model and empirical overview.
Figure 1. Conceptual model and empirical overview.
Jtaer 20 00138 g001
Jtaer 20 00138 g002
Figure 2. Relationship between expression certainty and consumer purchase behavior.
Figure 2. Relationship between expression certainty and consumer purchase behavior.
Jtaer 20 00138 g002
Jtaer 20 00138 g003
Figure 3. The mediation effects of perceived competence and perceived warmth.
Figure 3. The mediation effects of perceived competence and perceived warmth.
Jtaer 20 00138 g003
Jtaer 20 00138 g004
Figure 4. Mediation effect test across different ranges of expression certainty.
Figure 4. Mediation effect test across different ranges of expression certainty.
Jtaer 20 00138 g004
Table 1. Demographic details of the samples.
Table 1. Demographic details of the samples.
VariableMeasurementSamplesVariableMeasurementSamples
(Videos)
GenderMale76Product categoriesdaily necessities712
Female67baby products657
Number of fans<100,00041cosmetic552
100,000~1 million76clothing425
1 million~10 million21electronic product285
>10 million5food1211
Table 2. Negative binomial regression results for the impact of expression certainty on purchase behavior.
Table 2. Negative binomial regression results for the impact of expression certainty on purchase behavior.
AntecedentModel 1Model 2Model 3
β (se)pβ (se)pΒ (se)p
Expression certainty 0.241 0.0000.2040.000
Expression certainty 2 −0.1430.000
Gender0.014 0.7520.014 0.7420.014 0.714
Number of fans0.1410.0050.1320.0060.1440.002
Update frequency−0.0420.211−0.0330.201−0.0420.183
Product price−0.3370.001−0.3410.001−0.3120.001
Live-Stream Duration0.0210.8120.0220.8100.0240.721
Readability0.242 0.0040.255 0.0020.224 0.004
Information Volume0.0250.3710.0240.3620.0270.352
Monetary-Related
Information		0.257 0.0010.253 0.0010.2440.001
Social-Related
Information		0.2540.0020.2440.0030.2280.003
VIF1.471.473.95
Log-likelihood−6524.33−6442.37−6328.24
Wald χ2428.140.000431.950.000475.250.000
AIC5725.25528.15418.6
N384238423842
Notes: The superscript “2” represents the quadratic term of “Expression certainty” (i.e., Expression certainty squared).
Table 3. Results of robustness checks.
Table 3. Results of robustness checks.
Robustness of DataRobustness of the Dependent Variable
Model 4Model 5Model 6Model 7
Expression certainty0.17 **0.13 **0.23 ***0.19 ***
Expression certainty 2 −0.05 ** −0.12 **
ControlsYYYY
AIC4221.54105.63741.83605.4
Notes: ** p < 0.01, *** p < 0.001. The superscript “2” represents the quadratic term of “Expression certainty” (i.e., Expression certainty squared).
Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content.

Share and Cite

MDPI and ACS Style

Chen, Y.; Han, Q.; Lian, Z.; Zhu, W.; Jiang, Y. From Certainty to Doubt: The Impact of Streamer Expression Certainty on Consumer Purchase Behavior in Live-Stream E-Commerce. J. Theor. Appl. Electron. Commer. Res. 2025, 20, 138. https://doi.org/10.3390/jtaer20020138

AMA Style

Chen Y, Han Q, Lian Z, Zhu W, Jiang Y. From Certainty to Doubt: The Impact of Streamer Expression Certainty on Consumer Purchase Behavior in Live-Stream E-Commerce. Journal of Theoretical and Applied Electronic Commerce Research. 2025; 20(2):138. https://doi.org/10.3390/jtaer20020138

Chicago/Turabian Style

Chen, Yinjiao, Qianqian Han, Zhihua Lian, Weiming Zhu, and Yushi Jiang. 2025. "From Certainty to Doubt: The Impact of Streamer Expression Certainty on Consumer Purchase Behavior in Live-Stream E-Commerce" Journal of Theoretical and Applied Electronic Commerce Research 20, no. 2: 138. https://doi.org/10.3390/jtaer20020138

APA Style

Chen, Y., Han, Q., Lian, Z., Zhu, W., & Jiang, Y. (2025). From Certainty to Doubt: The Impact of Streamer Expression Certainty on Consumer Purchase Behavior in Live-Stream E-Commerce. Journal of Theoretical and Applied Electronic Commerce Research, 20(2), 138. https://doi.org/10.3390/jtaer20020138

Article Metrics

Back to TopTop