Examining direct sales as a violation of friendship expectations on WeChat

Abstract WeChat has emerged as the leading app in China, providing a platform for direct sellers. However, selling behaviors can be perceived as violations to friends, as WeChat was originally designed for private sharing. This study examines the relationships and communication dynamics of WeChat users when their friends become sellers, using the framework of Expectancy Violation Theory (EVT). The findings reveal that participants viewed their friends’ selling behaviors as moderately expected, low in importance, and moderate to highly negative in valence. The type and length of the relationship significantly influenced the perceived violation. Furthermore, participants reported changes in communication behaviors following their friends’ transition into sellers. This study sheds light on the impact of selling behaviors on WeChat friendships from an EVT perspective, offering insights into users’ perceptions and communication adjustments in response to these violations.

Technology and social media have emerged as the dominant platforms for conducting business (Ferrell et al., 2010).WeChat, the leading social media app in China, has played a significant role in driving the country's e-commerce expansion (Lau et al., 2015;Zhang, 2015).WeChat, a versatile social media application launched by Tencent in 2011, has captured the attention of Chinese internet users and amassed 1.082 billion monthly active users by the third quarter of 2018 (Tencent, 2018).Companies have created public accounts and subscription accounts, allowing WeChat users to explore promotions and discover new products.Notably, business accounts offering promotional information attract a considerable 41.9% of followers among all types of public accounts (WalktheChat, n.d).Within the WeChat community, a distinctive form of business has emerged known as direct sales.Direct sales involve a selling strategy where sellers establish direct contact with buyers (Luk et al., 1999).Unlike traditional retail experiences where buyers select products from stores, direct sales involve sellers personally delivering and introducing products to buyers.Typically, buyers in direct sales scenarios tend to be the sellers' friends, family members, and colleagues (Luk et al., 1999).

ABOUT THE AUTHOR
Xiaoti Fan is an assistant professor in the Department of Communication Arts at the University of Alabama in Huntsville.Her research explores relationship development and dissolution in online environments such as online dating sites, communication application, and social network sites.Additionally, she studies how deceptive communication impact personal relationships in multicultural contexts.
When WeChat arose as the most popular communication tool and social media platform in China, people started to directly sell products on WeChat through Moments as well as Groups (group chat function) to their friends and families.The products were mostly unbranded cosmetics and health supplements, and those businesses greatly profited (Li & Jin, 2016).When friendships meet businesses, conflicts often initiate (Grayson, 2007).Sellers on WeChat need to find a way to manage their personal relationships and their businesses spontaneously.WeChat is considered a private and mutual social networking app, similar to Facebook.Friendship on WeChat can be defined as anyone who sends a request to a user, and the user accepts the request (Bevan et al., 2014).One of the most popular functions of WeChat is Moments.Moments is a platform similar to Facebook timeline that is built into WeChat for users to post text, pictures, articles, and videos.Unlike Facebook timeline, Moments is only visible to WeChat friends.Approximately 61.4% of WeChat users reported checking their Moments every time they open WeChat (WalktheChat, 2016).
Although a significant amount of research has been conducted on direct sales models, marketing, and business successes on WeChat, few have studied the influences direct selling on WeChat has on relationships between sellers and their target buyers.This study aims to investigate the influence sellers have on their friendships on WeChat by employing Expectancy Violation Theory (EVT).Burgoon (1993) suggested that people all have expectations for others and their interactions; therefore, their expectations can be violated sometimes.The question this study asks is whether WeChat friends selling products is viewed as an expectancy violation by individual users.Since WeChat is considered a social networking platform and expected to be used for personal relationship building, doing business with a friend might seem unexpected.EVT has been studied on Facebook for unfriending behavior (Bevan et al., 2014).The three aspects of EVT examined include (1) expectedness (if certain behavior is expected); (2) importance (if the violation is considered important); and (3) valence (how negative or positive is this violation perceived) (Afifi & Metts, 1998;Bevan et al., 2014).
Since WeChat has become a part of Chinese people's daily routine (Lau et al., 2015;Zhang, 2015), it would be beneficial to learn whether the selling behavior of a friend on this platform, which was designed to enhance friendship, would hurt relationships instead.Moreover, the investigation might help sellers to better manage their friendships on WeChat.Therefore, this study is interested in exploring the behavior of direct selling on WeChat, and how individuals view it in terms of violation of the three aspects of EVT.

Expectancy Violation Theory (EVT)
The former EVT, Nonverbal Expectancy Violation Theory, was initially developed by Burgoon and Jones (1976) to explore the violation of personal distance.The original EVT has since been applied to studies on nonverbal communication violations (Burgoon & Hale, 1988;Burgoon et al., 1989).Burgoon (1993) further expanded the theory to investigate interpersonal verbal communication, defining expectancy as "an enduring pattern of anticipated behavior" (pp.31).Afifi and Metts (1998) defined the violation of such expectancies as "a behavior that a receiver notices as being different from the behavioral display that they anticipated" (pp.367).
According to EVT, perceived expectancy can be influenced by three factors: communicator characteristics, relationship characteristics, and context characteristics.In other words, the individual factors of the sender, such as demographics, appearance, and personality, along with the relationship between interactors in terms of familiarity, liking, and attraction, as well as the environmental situations such as privacy and formality, all play a role in shaping an individual's expectations of social interaction (Burgoon, 1993).Additionally, EVT argues that people's expectancies are understood and evaluated based on the effect of the reward valence (i.e., if the communicator makes the interaction rewarding).As a result, expectations of interaction are impacted by the reward valence, leading to various outcomes.Afifi and Metts (1998) argued that previous literature had a limited perspective by considering all violations as salient, significant, negative, uncertain, and surprising.They suggested that the valence of a violation could vary depending on the relationships and contexts involved (Afifi & Burgoon, 2000;Afifi & Metts, 1998).Therefore, Afifi and Metts (1998) conducted a study to measure different levels of expectedness, importance, and valence in individuals' perceptions of violations.The findings revealed various categories of violations within romantic partner communication, with these violations varying in their level of expectedness and importance.These three aspects of perceiving violations of expectancy made significant contributions and were subsequently tested in Bevan's (2003) study on sexual resistance messages and Bevan et al. (2014) study on Facebook unfriending.
Valence refers to the degree to which a particular violation is evaluated as positive or negative.Expectedness pertains to how different the violation is from the expected behaviors.Importance focuses on the severity of the impact a violation may have on the relationship.In terms of valence, EVT has been applied and expanded upon in various studies examining both nonverbal and verbal interpersonal communication.By manipulating reward valence, Burgoon et al. (1989) found that increased involvement indicated a positive meaning whereas decreased involvement was rated as a negative meaning.This suggests that the valence of a violation can be influenced by the level of reward experienced during the interaction.Additionally, Floyd and Voloudakis (1999) suggested that valence can be perceived differently depending on the direction of change in individuals' affectionate behaviors.
EVT has been widely extended to computer-mediated communication (CMC) contexts in numerous studies, demonstrating its applicability beyond face-to-face interactions (e.g., Burgoon et al., 2016;Ramirez & Wang, 2008).Modality switching suggested that short-term online associations led to more positive violations when meeting offline, whereas long-term online associations led to more negative violations (Ramirez & Wang, 2008).Furthermore, Burgoon et al. (2016) investigated human communication with embodied agents and examined the violations perceived by humans when interacting with computers.Moreover, Nicholls and Rice (2017) found that when expectancy violations were ambiguous to judge, the perception of such violations was mediated by the reward value associated with the communication.Several other studies have delved into popular social media platforms like Facebook (Bevan et al., 2014;McLaughlin & Vitak, 2011;Stutzman & Kramer-Duffield, 2010).However, the EVT literature lacks sufficient exploration of the most widely used social media app in China, WeChat.Given the enormous user base of WeChat (Tencent, 2018), it is crucial for communication scholars to expand EVT research to encompass this platform and gain a comprehensive understanding of expectancy violations within its unique context.

WeChat & business on WeChat
WeChat, a mobile application (app), launched in 2011 started as a free instant messaging app (Harwit, 2016).By offering over a dozen features, users can achieve different goals with the use of one app, which provided exceptional convenience to their mundane lives.The multifunctional app provides functions for communication (e.g., video chat, group messages), sharing (i.e., Moments), networking and/or dating (e.g., look around, shaking, drifted bottle), financial assistant (e.g., money transfer, red envelope, payments), and more (Harwit, 2016).By the end of 2011, WeChat has obtained 500 million users (Harwit, 2016), which grew to over one billion users by the third quarter of 2018 (Tencent, 2018).Not only did individual users enjoy the benefits WeChat brought to them, but the business also benefited from the public account and subscription function of WeChat (Yu et al., 2017).
A public account is a function that provides news media and businesses a platform to promote and communicate with a large range of users.Businesses use public accounts to promote new products, share news, and provide customer services (Yu et al., 2017).WeChat, as the most dominant social media app in China, has facilitated Chinese e-commerce growth in a significant way (Lau et al., 2015;Zhang, 2015).One special form of business arose among WeChat users using not the public feature, but the private feature of the app.Moments and instant messaging were found to be the most used features in WeChat, and individual users started to directly sell products on WeChat through Moments as well as through instant messaging individuals or groups of friends and families.The products sold were mostly cosmetics and health supplements, and those businesses greatly profited (Li & Jin, 2016).In Chinese culture, relationships hold immense significance as they are considered one of the fundamental pillars of people's lives.Traditionally, individuals are expected to form relationships based on kinship and geographic proximity (Luk et al., 1999).In this cultural context, relationships, known as Guanxi, hold great importance, and Chinese individuals are not inclined to easily turn down their families and friends (Merrilees & Miller, 1999).However, when friendships intersect with business dealings, conflicts often arise (Grayson, 2007).The dynamic interplay between personal relationships and business transactions requires sellers to navigate this delicate balance to maintain successful interactions on the platform.Finding strategies to harmonize personal relationships with business operations becomes crucial for sellers on WeChat.

Friendship
Previous research has extensively examined friendship dynamics in both online and offline contexts, with a specific focus on expectancy violations and their impact on social and media friendships.Social friendships were found to be less vulnerable to damage than media friendships caused by expectancy violations (Cohen, 2010).Moreover, expectancy violation by weak ties was found to be positively correlated to friends-only profiles on Facebook (Stutzman & Kramer-Duffield, 2010).More studies have explored online friendships in general.For instance, Mesch and Talmud (2006) investigated online social ties and found that adolescents seek social support from online friendships.Additionally, Chan and Cheng (2004) compared online and offline friendships and found offline friendships had greater interdependence, breadth, depth, code change, understanding, commitment, and network convergence; also qualities of cross-sex online friendships were higher than same-sex friendships (Chan & Cheng, 2004).Bevan et al. (2014) studied Facebook's unfriending behavior and its perceived expectancy violation by users being unfriended.They measured the valance, expectedness, and importance of such violation, as well as relational influence and usage intensity influence on the perceived violation.WeChat is a platform that allows users to dwell with close people and also encounter strangers (Wang et al., 2015).The definition of a WeChat friend, similar to Bevan et al. (2014) definition of Facebook, should be anyone who sent a request to a user, and the user accepts the request.

EVT and selling on WeChat
Since direct selling, especially selling on WeChat has been reviewed as inappropriate, this study aims to investigate whether WeChat users view the selling behavior of their friends as a violation of expectancies.Previous literature covered EVT in various offline and online contexts, but WeChat has not been studied under EVT as much.Most previous studies on WeChat focused on users' gratification and motives (Gan, 2017;Gan & Wang, 2015;Lien et al., 2017;Pang, 2016;Wang et al., 2015;Zhang et al., 2017), as well as the relationship between the state and WeChat (Harwit, 2016;Tu, 2016).Qiu et al. (2016) studied WeChat social messaging groups and how they have developed and evolved.Other studies conducted on direct selling on WeChat were mostly from business and marketing fields, which lack exploration in communication especially interpersonally.Therefore, this study aims to test the relationship factor of EVT and explore the influence this factor has on individual expectations.

RQ 1 :
To what extent is friends selling on WeChat viewed as an expectancy violation according to (a) valence (b) expectedness, and (c) importance?

Relationship type
As stated above, WeChat is a platform for both close friends and families, as well as strangers; thus, relationships vary to a major degree on WeChat.Users can only see posts from their WeChat friends and can only view comments from mutual friends.However, users can choose to block certain friends' posts or messages.Bevan (2003) stated that the type of relationship can affect how individuals evaluate a violation.McLaughlin and Vitak (2011) also indicated the closeness level of Facebook friends affects how users perceive expectancy violation.Most importantly, Bevan et al. (2014) found relationship closeness did, in fact, influence Facebook users' evaluation of the valence of the unfriending behavior as a violation.Thus, this study looks to investigate whether different types of relationships among WeChat users influence their perception of direct selling behaviors if they perceive it as an expectancy violation.
RQ 2 : How does the type of relationships influence the users' view of selling on WeChat as an expectancy violation according to (a) valence (b) expectedness, and (c) importance?

Length of friendship
While relationship types are a crucial factor in examining expectancy violation, the length of the relationship should also be considered.As mentioned earlier, WeChat friends can range from family members to strangers who connected online.Initially, WeChat users primarily added friends from their offline social network, but this dynamic has evolved over time.The length of relationships has been identified as a significant factor influencing the perception of violation (Bevan et al., 2014).Given that WeChat friends are not exclusively offline acquaintances, this study investigates both offline relationship duration and online WeChat friendship length in relation to expectancy violation.

WeChat usage intensity
Despite extensive research on various aspects of WeChat usage intensity, there remains a gap in understanding the impact of WeChat usage intensity on expectancy violations.In specific, researchers have explored the effects of WeChat usage intensity on user motivation and subjective well-being (Wen et al., 2016), as well as its influence on college students' well-being (Pang, 2018).Additionally, WeChat usage intensity has shown a positive association with social support (Wang et al., 2019).However, a significant gap exists in the literature concerning expectancy violations in relation to WeChat use intensity.While Bevan et al. (2014) investigated the relationship between Facebook usage intensity and views of unfriending behavior as a violation of three aspects, this study aims to explore a similar influence of WeChat usage on violations related to selling behavior.

Communication behaviors
Several studies have examined the consequences or aftermath of an expectancy violation.McLaughlin and Vitak (2011) argued that both the nature of the violation and the relationships between the communicators influence individuals' reactions to violations on Facebook.Specifically, minor negative violations by acquaintances were often ignored, while more significant violations could result in the deletion of Facebook friendships.;however, close friends who committed negative violations were often confronted.Moreover, Burgoon et al. (1989) found positive violations generally led to favorable outcomes whereas negative violations resulted in unfavorable outcomes.Furthermore, unexpected changes in affectionate behavior influenced recipients' perceptions of the violators' character (Floyd & Voloudakis, 1999).Lastly, Bachman and Guerrero (2006) proposed two types of communication following a violation: constructive (disclose feelings such as confrontation) and destructive (think about it alone or cry alone without disclosing the sender).This study aims to explore the communication changes that arise from expectancy violations related to selling on WeChat.
RQ 5 : What communication behaviors will individuals engage in after their friends become sellers on WeChat?
Participants have used WeChat for M = 6.9 (SD = 2.22) years.On average, participants have M = 449 WeChat friends (SD = 436.13).Participants spent M = 3.08 (SD = 2.97) hours on WeChat daily.Most of the participants used WeChat mainly for chatting or contact (40%), 15% used it for Moments, 8.3% used it for browsing news or articles, 3.3% used it for work, 1.7% used for paying or shopping, and 31.7%used for a combination of purposes.Almost all participants used the feature Moments (98.3%), and they browsed or posted on Moments 11.67 times a day on average (SD = 19.18).

Procedures
Participants accessed the survey from Sojump.com.The original survey was designed in English and was translated into Chinese for the target audience.The link was also distributed through the researcher's personal WeChat account.WeChat users read the recruitment messages on WeChat and clicked the link that took them to the survey.To be eligible, participants had to be at least 18 years, speak and read the Chinese language, and have access to a mobile device and WeChat software.Additionally, participants must have at least one WeChat friend who was a seller, and the participants were not sellers.Upon meeting eligibility criteria, participants read the approved Institutional Review Board (IRB) informed consent form and agreed to participate.The collection was anonymous and took participants an average of 6.72 (SD = 6.27) minutes to complete.Upon completion, each participant was paid 1 yuan (approximately $.15).The money was deposited into their WeChat wallet.
Initially, participants answered demographic questions including age, sex, and education level.Participants then answered questions about their general and specific WeChat usage including the length of usage, number of friends, hours of utilization, purposes of usage, and Moments' usage.Next, participants reported the number of friends they had on WeChat that were selling products and thought about one seller friend, and answered questions about this individual including their age, sex, selling strategies.Last, participants answered whether or not they followed or added this individual friend on other social media platforms.

Variables
Participants were then asked to identify the relationships they share with their seller friends.Relationship types included current and former family members, friends, and romantic partners, current and former classmates, co-workers, acquaintances, friends-of-friends, and those who only met online.Additionally, participants reported the length of their relationships both online and offline before their friends became sellers.Next participants completed a modified Expectancy Violation Scale (Afifi & Metts, 1998), where participants rated their level of viewing the selling behavior as an expectancy violation in terms of valence, expectedness, and importance.Participants then completed a modified Facebook Intensity Scale (Ellison et al., 2007).Exemplar items are discussed below.Last, they answered an open-ended question about any changes that appeared between them and their seller friends in terms of communication on WeChat.In the end, participants were thanked and paid.

Measurements
The current study utilizes two measurements adopted from Afifi and Metts (1998), and Ellison et al. (2007).

Aspects of expectancy violation
This measurement (Afifi & Metts, 1998) was originally designed to measure the expectancy violation expectedness, importance, valence, impacts on uncertainty, and attributions of such violation.The scale asks participants to think about one thing they have not expected and completed the items accordingly.Bevan et al. (2014) modified this scale to measure expectancy violation expectedness, importance, and valence on Facebook unfriending behavior using 7-point semantic differential scales with three items measuring each aspect.This study modified the prompt from Facebook unfriending to friend selling on WeChat with Bevan et al. (2014) differential scale.Two items measured violation expectedness (e.g., "She/him selling on WeChat was completely-not at all expected," M = 3.21, Cronbach's α = .82),and one item from Bevan et al. (2014) was dropped due to repetition.Three items measured violation importance (e.g., "This person selling on WeChat was a very unimportant-important relationship event," M = 2.33, Cronbach's α = .87).Three items measured violation valence (e.g., "This person selling on WeChat made me feel like they care about our relationship a lot-didn't care about our relationship at all," M = 4.06, Cronbach's α = .81).Higher values indicated a higher level of violation expectedness, importance, and negative valence.Ellison et al. (2007) created this scale to measure the level of integration of Facebook into participants' everyday lives.This study adopted the 5-point Likert scale (1 = Strongly disagree, 5 = Strongly agree) with minor modifications (i.e., changing Facebook to WeChat in the instruction).Six items were adopted from the Facebook intensity scale (M = 4.06, Cronbach's α = .75).Higher values indicated a greater intensity of WeChat usage.

Results
Participants were asked to think about one friend who sells on WeChat and answered questions about this friend.As shown in Tables 3 and 4, on average, participants reported having 8.23 friends who engaged in selling behaviors (SD = 8.84).Participants' seller friends were 86.7% female, and their average age is M = 28.27(SD = 4.46).98.3% of the sellers post on Moments of their products, 25% send direct messages to their contacts, 50% send messages in their group chat rooms, and 5% use other ways to sell.Only 36.7% of the participants added or followed their seller friends on other social media.

Perceived expectancy violation
RQ 1 examined the extent to which WeChat users view their friends selling the product as an expectancy violation in terms of expectedness, importance, and valence.On a scale of 7, participants reported viewing their friends selling on WeChat as moderately expected (M = 3.21, SD = 1.74,Med = 3.00), low in importance (M = 2.33, SD = 1.49,Med = 2.00), and moderate to highly negative in valence (M = 4.06, SD = 1.22,Med = 4.00).

Relationship closeness
RQ 2 asked how the types of relationships influence the violation.Participants answered questions about one friend who sells on WeChat and identify their relationship (See Table 5 for the list of relationships).To determine if there was a significant difference in relationship types and users' views of the violation (expectedness, importance, and valence), a one-way Multivariant Analysis of Variance (MANOVA) was calculated using the type of the relationship as the independent variable, and evaluation of violation as the dependent variable.A significant main effect was found for expectancy violation: Wilks' k = .42,F(3, 27) = 1.79, p = .016.Moreover, significant univariate main effects were found for expectedness F(1, 9) = 3.68, p = .001,η 2 = .29,but not for importance F(1, 9) = 1.55, p = .157,η 2 = .08and valence F(1, 9) = 1.43, p = .200,η 2 = .06.

Length of relationship
RQ 3 assessed whether the length of the relationship influenced users' perception of the violation.Participants identified how long (in years) they have known their friends before he/she became a WeChat seller (M = 4.99, SD = 7.17).They also indicated how long (in years) they have been WeChat friends before he/she became a seller (M = 2.74, SD = 4.83).A moderate negative relationship was found between valence and relationship length off WeChat, r (60) = -.30,p < 0.05.No significant correlation was found between the other pairs (See Table 6 for a complete correlation between the length of relationships and aspects of violation).

WeChat usage intensity
RQ 4 explored how the intensity of using WeChat would influence users' views on the violation.The correlation test suggests there were no significant relationships found between the usage intensity and expectedness (r (60) = −.09),importance (r (60) = −.12), and valence (r (60) = −.04).See Table 7 for correlations between usage intensity and all three aspects.

Communication changes
RQ 5 investigated specific changes in communication among WeChat users after they experienced expectancy violations.Four themes were generated from the open-ended response using analytic induction to generate themes from the only available data due to the nature of a small set of cases (Bulmer, 1979).The guide is directed at discovering important themes and tests for commonality.The researcher read through the responses and generated four major themes: 1) no change in behaviors, 2) reduced communication, 3) increased communication, and 4) conditional communication.
The first theme is no change in behaviors, which indicated the same communication as they usually engage with their friends (56.7%).Examples of this theme include "none," "no change, already a seller when we met."The second theme is reduced communication (28.3%), which is defined by less communication engagement than usual with their friends.This type of change involves either less direct messaging, blocking Moments, or blocking friends.Examples include "reduced the chance for us to know each other better," and "I blocked her Moments."The third theme is increased communication (5%), which means users communicated more to their friends after they became sellers.A couple of examples are "paying more attention to her Moments to see if I need to buy anything from her," and "I communicate with her more to learn about discounts and shopping skills."The last theme is conditional communication (10%), which showed by users who talk to their friends only when they are in need.Examples of this include "Our communication decreased outside of shopping, but when I need to buy something, I will contact her," and "I occasionally contact her about a product."Lastly, when participants were asked about the additional information they wanted to provide, some interesting comments were recorded.One participant said, "too many sellers indeed are not beneficial to the development of WeChat communication environment."Another said, "if a friend I'm not familiar with became a seller, I would block his/her Moments."One shared, "She sells jewelry but also has a store offline.She often posts on Moments about her product, but nothing else." Lastly, someone said, "Sellers always recommend blindly just for the sake of selling, they send ads all the time.I personally loathe these unbranded and unhealthy products."

Perceived violation
The findings of this study indicate that WeChat users perceive their friends' selling behaviors as a violation of their friendships.However, participants generally considered the violation to be moderately unexpected, of moderate to high negativity, but not of significant importance.While a few users expressed extremely negative comments towards direct sellers on WeChat, the majority of users showed indifference towards this behavior.Sellers predominantly utilized the strategy of posting on Moments, which may explain why users were less concerned.As users rarely received individual or direct messages from these sellers, they did not perceive themselves as targeted buyers.Consequently, a large portion of users did not alter their communication behaviors in response to the violations.This finding aligns with the assumption of EVT, which suggests that the consequences of a violation are determined by individuals' evaluations of the violation.

Relationship and violation
This study results aligned partially with previous research (Afifi & Burgoon, 2000;Afifi & Metts, 1998;Bevan et al., 2014;McLaughlin & Vitak, 2011) and supported the assumption of EVT that relationships influence violations (Burgoon, 1993).The type of relationship has been found to impact expectedness, but not the importance or valence of the violation.Bevan et al. (2014) classification of relationship closeness/ties may not fully apply to the WeChat user population.
Close-ties and weak-ties were categorized based on relationship types such as family, coworkers, and acquaintances.However, in WeChat, individuals often have multiple types of complex relationships.For example, some individuals may have family members or relatives with whom they share no significant resources or rarely interact, while others may have a coworker (classified as a weak-tie) who is also a close friend (classified as a strong-tie).Therefore, this study refrained from categorizing ties and instead examined all 13 types of relationships present on WeChat.
Relationship length off WeChat was tested to be negatively correlated with the valence of the violation.Relationship length serves as an indicator of relationship closeness and ties, suggesting that the longer the relationship, the closer the individuals involved.The results indicated that longer friendships were associated with a less negative perception of selling behaviors.However, the study did not find a similar correlation between expectedness and importance.Despite having longer relationships, WeChat users did not view their friends' selling behavior as more unexpected or consider it more important.This finding can be explained by cultural influences, such as the desire to maintain connections with kinships (Luk et al., 1999) or the fear of losing connections within the social network (Merrilees & Miller, 1999) that have been nurtured over an extended period of time.

Intensity of WeChat usage
Interestingly, the study found no correlation between the intensity of WeChat usage and the three aspects of perceived violations, which deviates from the findings of Bevan et al. (2014).This result suggests that the intensity of WeChat usage does not influence users' perspectives on their friends' selling behavior.This can be attributed to the selling strategies employed.When sellers post in Moments, users may not pay close attention to these messages and may even block or skip them altogether.Consequently, the intensity of usage may not significantly affect what users perceive.Posting in Moments can be seen as a form of mass communication rather than interpersonal communication.In contrast, previous studies examining WeChat usage intensity have primarily focused on one-on-one interactions (e.g., Bevan et al., 2014;Wang et al., 2019).

Communication behavioral change
The qualitative analysis revealed four major themes regarding users' communication behavioral changes with their seller friends: 1) no change in behaviors, 2) reduced communication, 3) increased communication, and 4) conditional communication.The majority of users reported no significant changes in their communication patterns, maintaining contact only when necessary.A considerable number of users mentioned reduced communication, while some participants reported completely blocking their seller friends.These findings reinforce the notion that most users do not perceive selling behavior as an important violation and consequently do not take action in response to the violation.Bevan et al. (2014) highlighted the significance of perceived violation as a predictor for contacting a former friend regarding unfriending, compared to valence and expectedness.The results of this study can be attributed to users' perception that selling on WeChat is neither important nor unexpected.However, it is important to consider the influence of cultural values.Many participants reported having weak relationships with their seller friends, such as those formed online or former co-workers (see Table 5).These results align with the findings of McLaughlin and Vitak (2011), where negative violations from acquaintances were often disregarded.Furthermore, participants in this study predominantly engaged in destructive communication following violations (Bachman & Guerrero, 2006).

Implications
This study extends EVT into a novel area of focus by examining friendships on WeChat and exploring the influence of relationship factors on violation consequences within the EVT model.Additionally, the investigation of WeChat selling behaviors offers valuable insights into Eastern cultural perspectives within the realm of mediated interpersonal communication scholarship, as well as the broader field of interpersonal EVT research.The study sheds light on a relatively unexplored social media platform within the EVT framework, uncovering both consistent and inconsistent findings in relation to previous literature.
More importantly, this study provides practical recommendations for both WeChat users who are sellers and those who are not, aiming to foster better communication between the two parties by acknowledging general users' perspectives on selling behavior.The findings indicate that longer relationships between sellers and users tend to elicit more positive reactions when a friend becomes a seller.For sellers who have just met someone, it is advisable not to immediately promote products, as it is more likely to be perceived as a negative violation.Furthermore, relationship types play a role in individuals' perceptions, suggesting that sellers should employ different strategies targeting different audiences.Additionally, the intensity of WeChat usage does not significantly influence users' surprise when encountering selling posts, allowing sellers to approach individuals from different generations in a similar manner.Importantly, the majority of users do not take negative actions towards the seller, suggesting that it is possible to engage in selling on WeChat while maintaining good relationships with friends and minimizing the risk of losing those friendships.

Limitations & future directions
The data collection process for this study posed significant challenges, resulting in a smaller usable sample size than anticipated.Participants displayed limited motivation to complete the survey, leading to conflicting and unusable data (e.g., random or exaggerated responses such as claiming usage of WeChat 1000 times a day).
While this study delved into the relationship factor within EVT in the context of WeChat selling behaviors, there is scope for exploring individual factors and contextual factors in future research.How does the valence of rewards influence the three aspects of violation?Additionally, investigating the perspective of sellers within EVT's assumptions could shed light on how sellers view the behaviors of their peers.What expectations might be violated when sellers attempt to sell products to one another or engage in promotional exchanges?Moreover, it is encouraged to conduct cross-cultural and cross-media studies, incorporating platforms such as Facebook and WeChat, to better understand expectancy violation within diverse cultural contexts.This approach would provide valuable insights into the multicultural aspects (context factors) inherent in the study of EVT.

Conclusion
WeChat has become deeply ingrained in the daily lives of Chinese internet users (Harwit, 2016;Yu et al., 2017), with many of their friends engaging in product selling on this platform (Yu et al., 2017).However, the concept of expectancy violation has been overlooked in the context of WeChat interpersonal communication.This study sheds light on the role of EVT in the communication between general users and their seller friends on WeChat.The findings offer implications for EVT and provide relationship management strategies for WeChat sellers.Notably, as most users do not alter their communication patterns following a violation, sellers need not be overly concerned about jeopardizing interpersonal connections while engaging in selling activities.

RQ 3 :
How does the length of friendships influence users' view of selling on WeChat as an expectancy violation according to (a) valence (b) expectedness, and (c) importance?

RQ 4 :
How does the WeChat usage intensity influence users' view of selling on WeChat as an expectancy violation according to (a) valence (b) expectedness, and (c) importance?

Table 6 . Pearson Correlations among length of friendships on/off WeChat and violation Friendship off WeChat Friendship on WeChat Expectedness Importance Valence
**Correlation is significant at the 0.01 level (2-tailed).*Correlation is significant at the 0.05 level (2-tailed).