Effects of WeChat on the Closeness between Chinese
Parents and Their Children in Emerging Adulthood
Author Note
Xiaoyang Zhang, Student-ID: 10389792, Master’s Thesis, Graduate School of
Communication, Master’s programme Communication Science.
Abstract
This study aimed to examine the associations between WeChat (including its text/voice messaging and social networking service, Moments) communication and the closeness between Chinese parents and their children in emerging adulthood, as well as moderating effects of trait-like self-disclosure, geographic distance and gender of parents. A survey of 195 Chinese adults between 18 to 25 years revealed that child-parent closeness did not predict how frequently emerging adults communicated with their parents via WeChat, while communication via WeChat messaging had a positive effect on both child-mother and
child-father closeness. Besides, communication with fathers via Moments positively affected the extent of closeness that emerging adults felt with their fathers. Child-mother closeness, however, was not influenced by their communication via Moments. Consist with media rich theory, communication via WeChat messaging had a greater effect on child-parent closeness compared to communication via Moments. Moreover, the effects of WeChat communication on child-parent closeness were not moderated by emerging adults' self-disclosure, geographic distance between emerging adults and their parents, or gender of parents. Limitations and practical implications are discussed.
Effects of WeChat on the Closeness between Chinese Parents and Their Children in Emerging Adulthood
Emerging adults, that is, individuals who are aged 18 to 25, are the main users of the Internet in China. Among 600 million Internet users and 500 million mobile Internet users, people in their twenties account for 30%, which is the biggest age group (CNINC, 2014). Emerging adults were born between 1989 and 1996, and are often identified as “digital natives” and “the Post-90s generation”. They are heavy users of the latest communication technologies, such as instant messaging, e-mail, and social networking sites, which have become part of their daily life (Boyd & Ellison, 2007; Subrahmanyam & Greenfield, 2008). In China, “the Post-90s generation” are raised as only children and spoiled by their parents. The consequence is that most emerging adults are still financially and psychologically
dependent on their parents, even after moving out of parental houses (Arnett, 2004; Shi, 2009). They are interested in communicating with their parents via phone calls and other
technologies in order to feel at home (Shi, 2009). In this case, communication technologies compensate for the lack of face-to-face conversation and greatly help children keep in touch with their parents (e.g., Chen & Katz, 2009; Licoppe, 2002).
Unlike young people who quickly become handy with electronic media, most of their parents born in the 60s and 70s have difficulties using the latest technologies. One of the biggest obstacles is illiteracy or limited pinyin skills (Chinese phonetic system used to enter Chinese characters into electronic devices). Even though some parents learn to use popular communication technologies, such as mirco-blogging, to reach their adult children,
Remarkably, a new technology has built a bridge for interaction between Chinese emerging adults and their parents in recent years. It is an Internet-based mobile
communication and social networking application called WeChat (Weixin in Chinese). Initially, WeChat enabled users to easily send text, voice, image, and video messages, but gradually added more services such as social networking and online games. Owing to the usability of WeChat’s basic functions (e.g., voice-based messaging does not require pinyin skills), it is not only heavily used by the younger generation, but it is also a convenient commination tool for their middle-aged parents (CNNIC, 2013). This is evident in the great number of WeChat users - 468 million (Millward, 2014).
Despite emerging adults and their parents’ extensive use of WeChat, studies have yet to consider the effect of WeChat use on child-parent closeness. This lack of research is remarkable because maintaining close child-parent relationships in emerging adulthood still plays an important role in children’s healthy emotional and social development (Liu, 2006). Hence, the present study aims to fill this gap by investigating whether communication between emerging adults and their parents via WeChat affects their closeness.
Theoretical Background WeChat as A Communication Device
WeChat is an Internet-based mobile application that allows users to connect with friends via different communication channels. Among all the channels, text and voice messaging are most widely used by 90% of users, followed by the Moments social networking service (CNNIC, 2013).
Text and voice messaging. WeChat text messaging allows users to send text
information after entering characters. Besides the traditional Chinese input method (pinyin), WeChat provides speech recognition that translates spoken words into text. WeChat voice messaging allows users to quickly record a message by pressing the "Hold to Talk" button, the release of which will send the message. Both text and voice messaging are easy to use, even for people who are illiterate or unable to type Chinese in pinyin. This service is free of charge.
Social networking service Moments. Like all social media, Moments allows users to create and share content with their WeChat friends. It is similar to Facebook where users can share their life stream, feelings, opinions, and experiences in words and pictures. Also, it is possible to like and comment on statuses shared by their contacts. Users can set visibility to certain friends for each ‘status’, which efficiently bypasses their concerns about privacy. Moreover, users are able to set limitations to avoid seeing statuses by certain contacts. Again, this service is free of charge.
Associations between Child-Parent Closeness and Communication
The current study aims to examine the association between WeChat use and child-parent closeness. The use of WeChat is defined as the combination of frequency and extent to which emerging adults use its voice and text messaging functions, and its social networking service, Moments. Child-parent closeness refers to the psychological closeness a child feels towards his or her biological parents (Golish, 2002). Homans’ system of
interpersonal relations and social exchange theory provides an important framework for this study as it elaborates on the relationship between social interaction and affection.
According to Homans’ system of interpersonal relations (1950), positive sentiment between two people makes them prone to interact with each other, and the understanding gained from these interactions, in turn, stimulates more positive sentiment. Thus, individuals tend to interact with people they like, and the frequent interaction renders them closer
compared to others who have less contact. Children who have a strong bond with their parents are predicted to contact them more frequently. They may acquire greater understanding and support from their parents, which strengthens the child-parent closeness. As Lawton,
Silverstein and Bengtson (1994) put it: “Those who share common experiences are likely to develop a collective identity and a sense of shared meanings that build empathy in the relationship; simultaneously, positive sentiments gained from the relationship serve as symbolic rewards for maintaining or increasing the on-going interaction” (p. 58).
Homan’s social exchange theory denotes that individuals prefer to repeat the behaviour (e.g., communication) if it produces positive outcomes (Homans, 1958). This proposition can explain why children are willing to frequently communicate with their parents if they gain benefits such as happiness, understanding and love during the interaction.
Homan’s social exchange theory has been successfully applied to examine the relationships between adult children and their parents in previous studies (Parikh, 2009). The theory has been further corroborated by the finding that interaction and affection between children and their mothers mutually reinforce each other (Lawton et al., 1994).
According to Homan’s theories regarding reciprocal interaction between
association between communication via WeChat and closeness between Chinese parents and their children in emerging adulthood.
Effects of Closeness on WeChat Communication
Child-parent closeness refers to the emotional closeness a child feels towards his or her parents (Golish, 2002), which pre-exists and may influence media use (Valkenburg & Peter, 2013). In accordance with Homan’s system of interpersonal relations (1950), the closer the relationship between emerging adults and their parents, the greater their need to interact with one another. When emerging adults move away from their parental house,
communication via electronic devices becomes the main form of interaction (Shi, 2009). With the increasing popularity of WeChat across younger and older generations, WeChat,
especially its messaging and social networking service Moments, becomes part of their daily communication. Several studies have found a positive association between child-parent relationship and electronic communication (Chen & Katz, 2009; Geser, 2005; Lee, 2002). Additionally, a directional positive relationship from child-mother closeness to contact frequency (visits and phone calls) has been proven by Lawton et al. (2014). Accordingly, there is a reason to believe that the closeness between emerging adults and their parents positively affects their communication via WeChat. The following hypotheses are formulated:
H1: The closer the child-mother relationship that emerging adults feel, the more emerging adults use WeChat including (a) text/voice messaging and (b) Moments to communicate with their mothers.
H2: The closer the child-father relationship that emerging adults feel, the more emerging adults use WeChat including (a) text/voice messaging and (b) Moments to communicate with their fathers.
Effects of WeChat Communication on Closeness
Empirical findings also corroborate Homan’s theoretical propositions regarding the positive effects of communication on affection. Specifically, some studies have demonstrated that interactions via phone calls benefit interpersonal relationships (Gentzler, Oberhauser, Westerman, & Nadorff, 2011; Licoppe, 2004). For instance, Gentzler et al. (2011) showed that college students who reported more frequent phone communication had a better relationship with their parent(s). Furthermore, messaging encourages communication and sharing within a household because of the reduced coordination and scheduling costs (Yarosh, Davis, Modlitba, Skov, & Vetere 2009; Markopoulos et al., 2004). Users can administer the communication according to their own schedule by choosing when to read and reply
messages (Sponcil & Gitimu, 2012). Likewise, other forms of electronic communication (e.g., instant messaging) facilitate close relationship maintenance, especially for long-distance relationships (Lee, 2003).
Given that the above-mentioned forms of electronic communication benefit intimate relationships including child-parent closeness, it is likely that WeChat communication may also increase the closeness that a child feels towards his or her parents.
WeChat’s social networking service, Moments, is a good platform for staying updated with each other. On the one hand, due to the socialization effect (Trepte & Reinecke, 2013), users tend to disclose their daily life, feelings and thoughts by writing their “status” on social
media. On the other hand, users are able to conduct either one-sided communication by viewing people’s “statuses” or two-way interaction through commenting and replying (Sponcil & Gitimu, 2012). The reciprocity on Moments may help children and their parents enhance familiarity, increase mutual understanding, and even cultivate common opinions.
Unlike many daily updates on social media, people usually use messaging to arrange face-to-face activities, signal one’s availability, and discuss issues (Kim, H., Kim, G., Park & Rice, 2007). Although some of the mentioned studies indicated a positive effect of messaging communication on interpersonal relationships, a few studies failed to find this association (Jin & Pena, 2010; Gentzler et al., 2011). This might be due to the limited amount of information carried by one text message, or because voice is not available during the communication (Jin & Pena, 2010). These two potential disadvantages are overcome by WeChat’s voice
messaging feature that enables people to send a voice-based message at no cost. Not only does the free-of-charge property of voice messaging encourage longer communication periods, but also does it help to reduce ambiguity and uncertainty often related to text messaging, due to better information flow and the presence of intonation (Daft & Lengel, 1986; Wang & McDonald, 2013). Moreover, creation and use of a family group on WeChat allows group conversations among family members and helps cultivate a sense of community and connection (Church & de Oliveira, 2013). WeChat messaging is thus supposed to perform better than traditional messaging. Accordingly, the current study proposes that WeChat text and voice messaging positively affect child-parent closeness as well.
H3: The more emerging adults use WeChat including (a) text/voice messaging and (b) Moments to communicate with their mothers, the closer the child-mother relationship that emerging adults report.
H4: The more emerging adults use WeChat including (a) text/voice messaging and (b) Moments to communicate with their fathers, the closer the child-father relationship that emerging adults report.
Following from the hypotheses, a model showing the relationships among
child-mother closeness, child-father closeness and communication via WeChat was drawn in Figure 1.
Figure 1. Hypothesized model on the relationships among child-mother closeness, child-father closeness and communication via WeChat.
Different Effects between WeChat Messaging and Moments on Closeness
According to the above discussion, WeChat text/voice messaging and its social media service Moments have different features and affordances. WeChat messaging is often used to “have conversation”, while Moments allows staying updated with each other. Besides, they differ in the degree of richness they process: messaging has a higher ability to reproduce the transmitted information than Moments. For instance, tone of voice can be reproduced by voice messaging but not Moments. In line with media richness theory (Daft & Lengel, 1986),
richer communication media can transmit more information and help improve understanding more than less rich communication media, leading to a more effective communication and a bigger media influence. Therefore, communication via WeChat messaging is supposed to have a greater effect on child-parent closeness than communication on Moments.
H5: Communication with mothers via WeChat messaging has a greater positive effect on child-mother closeness than that via Moments.
H6: Communication with fathers via WeChat messaging has a greater positive effect on child-father closeness than that via Moments.
Moderators of the Association between Mediated Communication and Closeness
Current media effect models emphasize the importance of considering the role of individual differences in media effects. For example, differential susceptibility to media effects model (DSMM) states that individual differences such as personality may enhance or reduce media effects on cognitions, emotions, attitudes, beliefs, physiology, and behaviour (Valkenburg & Peter, 2013). This explains why certain individuals are more susceptible to specific media than others. Below, specific factors will be discussed to determine how WeChat use may affect child-parent closeness in contrasting ways for different people.
Self-Disclosure. Self-disclosure is defined as an individual’s verbal or nonverbal
admission of their personal information, such as feelings, opinions, likes and dislikes, to others (Greene, Derlega, & Mathews, 2006). Self-disclosure is not only about daily life information (e.g., I drink coffee every day), but also concerns private personal information that is not easily shown to others, such as secrets, concerns, fears, and failures (Greene et al., 2006). Altman and Taylor’s (1973) social penetration theory asserts that self-disclosure
consists of two dimensions: breadth and depth. Breadth of self-disclosure indicates the range of discussed topics in the communication. Depth of self-disclosure refers to the intimacy level of communication; that is, the degree to which private or personal information is revealed.
Many studies have demonstrated that people’s willingness (i.e., the extent to which individuals would like to self-disclose) and ability to reveal their feelings and thoughts to others play a vital role in the development and maintenance of close relationships. Disclosure of both daily life information and privacy, as well as breadth and depth of self-disclosure are important (Altman & Taylor, 197; Greene et al., 2006). In previous studies, self-disclosure was seen as either a personality-like trait (e.g., Archer, 1979) or a particular behavior during interpersonal interaction (e.g., Solano, Batten, & Parish, 1982). The present study regards self-disclosure as a trait, which pre-exists WeChat use. Following the DSMM (Valkenburg & Peter, 2013), emerging adults’ self-disclosure as one dimension of dispositional susceptibility is expected to moderate the positive effect of WeChat use (incl. text/voice messaging and Moments) on child-parent closeness. Specifically, if emerging adults are prone to disclose themselves to parents during the communication on WeChat, this communication may be more efficient in increasing the feeling of closeness to parents. So, the following is expected:
H7: The positive effect of WeChat use including (a) text/voice messaging and (b) Moments on child-mother closeness is stronger for emerging adults with higher levels of self-disclosure tendencies than those with lower levels of self-disclosure tendencies.
H8: The positive effect of WeChat use including (a) text/voice messaging and (b) Moments on child-father closeness is stronger for emerging adults with higher levels of self-disclosure tendencies than those with lower levels of self-disclosure tendencies.
Geographic distance. Geographic distance refers to the physical distance between emerging adults and their parents. Studies have shown that adult children who live within one hour from their parents are inclined to feel close to parents through frequent contact, while fewer visits occur when children live further away from their parents (e.g., Lawton,
Silverstein & Bengtson, 2014). Geographic remoteness weakens emotional closeness between children and parents due to reduced opportunities for face-to-face communication (Licoppe, 2002). When children live far away from their parents, face-to-face communication will become too costly to remain a regular activity (Lee, 2002). Also, the cost of long-distance telephone calls will run up, which may lead to less communication. In this situation, mediated communication, especially low-cost online communication, becomes more relevant to
maintain a long-distance close relationship. This was supported by a longitudinal study, which found that face-to-face and telephone communication decreased, whereas interaction through e-mails and instant messaging increased after participants moved away from their co-resident families and partners (Lee, 2002). Consequently, there will be increasing effects of WeChat communication on child-parent closeness when they make less visits and phone calls due to growing time and economic cost in a long-distance relationship. Thus, the present study assumes that emerging adults feel closer to their parents when they use WeChat more often to contact them, and this association is stronger for emerging adults who live further away from their parents compared to those who live closer to their parents.
H9: The positive effect of WeChat use including (a) text/voice messaging and (b) Moments on child-mother closeness is stronger for emerging adults who live further away from their mothers than those who live closer to their mothers.
H10: The positive effect of WeChat use including (a) text/voice messaging and (b) Moments on child-father closeness is stronger for emerging adults who live further away from fathers than those who live closer to their fathers.
Parents’ Gender. It is widely acknowledged that adult daughters and mothers are
more likely to feel emotionally attached compared with adult sons and fathers. This common view is demonstrated by academic research as well (e.g., Lye, 1996). It might be because females feel more obliged and expressive to maintain family relations and thus are more active in family relationship maintenance than males (Lawton, 1994). Some empirical research found that mothers are central to the maintenance of family ties by providing an information exchange between family members (e.g., Chen & Katz, 2009; Sawhney and Gomez, 2000). If mothers cannot meet their children in daily life, they more actively keep in touch with their children far away in different ways, such as by mobile phones, compared to fathers (Sawhney & Gomez, 2000). Given that communication between mothers and their children is richer and there are more diverse contact approaches, the role of WeChat should be weaker for child-mother closeness than child-father closeness. That is to say, communication through WeChat is particularly valued for the closeness between children and fathers.
Therefore, the study devises the following assumption:
H11: The positive effect of WeChat use including (a) text/voice messaging and (b) Moments on child-parent closeness is stronger for child-father relationship than child-mother relationship.
The Role of Alternative Communication Channels
Nowadays, multiple forms of communication are often used by the young generation to contact their parents (e.g., face-to-face, phone calls and texts), and the combined effect of this varied communication helps maintain the child-parent relationship. There are six main contact channels including face-to-face conversation, voice call (incl. phone and Internet calls), voice messaging, video chats, text messaging (incl. phone text and instant messaging), and interaction on social media adopted by Chinese children and parents (Shen, 2014). Interaction through these channels will be controlled when the study tests the effect of WeChat communication on child-parent closeness.
Moreover, parents’ divorce, as well as adult children’s marriage can affect
child-parent closeness. Parents’ divorce potentially leads to alienation of children as shown in greater physical and psychological distance (Aquilino, 1994). It is more likely for married children to understand their parents and consequently have closer ties than unmarried children (Lawton, 1994). Thus, children and parents’ marital status are supposed to be controlled too.
Summary
All in all, the study proposes reciprocal positive effects between child-parent closeness and communication via WeChat messaging and Moments, for that concerning both mothers and fathers alike. Besides, it is suggested that communication with parents via WeChat messaging has a greater positive effect on their closeness than that via Moments. Furthermore, it is predicted that the strength of the effects of WeChat on child-parent closeness depends on certain boundary conditions. Specifically, the positive effects of WeChat messaging and Moments on child-parent closeness are stronger for emerging adults
with higher self-disclosure tendencies than those with lower self-disclosure tendencies, stronger for emerging adults who live further away from their parents than those who live closer to their parents, stronger for child-father relationship than child-mother relationship.
Method Participants
A non-probability convenience sample was recruited via WeChat and Sina Weibo (a major Chinese micro blogging service). This sampling procedure seemed appropriate given limited resources and a tight time schedule. Only participants who (1) had fully completed the survey, (2) were of Chinese nationality, and (3) were between 18 and 25 years were included in the study. Also, only participants who did not live in the same household as their biological parents were included in the analyses, because the study aimed to investigate how WeChat functioned in a long-distance relationship between family members. 195 valid participants who had both father and mother were obtained, with 67.2% female and 32.8% male. On average, participants were 23 years old (SD = 1.84). 46.2% had a Bachelor’s degree and 48.7% obtained a Master’s degree. Participants come from 21 provinces, 3 municipalities, and 1 special administrative region (Hongkong) with the most represented regions being Sichuan (22.1%), Beijing (16.9%), and Guangdong (6.7%). All other regions were summarized as East China (18.4%), South Central China (15.8%), North China (6.6%), Northeast China (5.6%), Northwest China (4.1%) and Southwest China (3.6%).
Procedure
A self-administered online survey was designed in English and translated into
conducted before the survey was officially launched. According to feedback from the
pre-tests, final changes were made and the survey was distributed on WeChat and Sina Weibo by an anonymous link. In total, the survey consisted of 29 questions and took approximately 10 to 15 minutes to complete. The survey was closed after three weeks of data collection.
A self-administered online survey was used for three reasons. First, emerging adults frequent the Internet, so it is easy for them to get access to an online survey. Second, the child-parent closeness is a private issue for many people, and a self-administered survey can well protect participants’ privacy. Third, an online survey is cost-effective and timesaving.
The survey comprised seven parts. The first part contained the cover story explaining that the researcher was a student at the University of Amsterdam conducting research in the use of electronic communication within Chinese families as part-requirement for her Master’s degree and that the study welcomed the participation of people with a Chinese nationality who were aged 18 to 25. Participants were further told that their total anonymity was
protected and they were free to stop at any time. Participants’ informed consent was obtained in this part too.
The second part asked questions on participants’ age, nationality, whether their mother and father alive, and living distance from their parents. These questions were presented in the beginning of the survey not only because they were very important variables, but also it allowed participants to skip relevant questions accordingly. The third parts measured the frequency of child-mother and child-father communication via face-to-face communication, voice calls, video chats, text messaging, voice messaging and social media in daily life. Communication via WeChat messaging and Moments was assessed in part four. Participants
were asked to indicate how close they felt to their mothers and fathers in the fifth part. The sixth parts measured participants’ level of self-disclosure to mothers and fathers. The last part was related to demographic data, including gender, educational background, hometown, participants’ and their parents’ marital status, and the family’s standard of living. Participants were thanked for their participation after submitting the survey.
During the process, participants were not allowed to go back to previous pages to redo questions given potential effects of child-parent closeness questions on their answers to communication frequency. They were also forced to give an answer to each question, but the option "I do not know" was provided for possibly sensitive questions.
Measures
Child-parent closeness. Participants’ perceptions of closeness to their mothers and
fathers were separately measured by the 10-item Closeness to Parents Scale (Buchanan, Maccoby, & Dornbusch, 1991). The Closeness to Parents Scale has been successfully employed to measure Chinese parent-adolescent relationship in a previous study (Liu, Fang, Zhou, Zhang & Deng, 2013), with a Cronbach’s α of 0.91 for mother-adolescent relationship and 0.91 for the father-adolescent relationship. To fit Chinese language habits, the wording was simplified in the current study using statements (e.g., “My mother knows what I am really like”) and a 5-point Likert scale (1 = strongly disagree to 5 = strongly agree), rather than the originally used questions (e.g., “How well does your mother know what you are really like?”) and the 5-point scale (1 = not at all to 5 = very). Example items included, “I feel close to my mother” and “I feel confident that my mother would help me if I had a problem”. A bivariate correlation analysis showed that item 3 “I feel that I have to be careful about what
I say to my mother” and item 10“I would not feel uncomfortable to ask my mother for money if I needed it” had no or small correlations (r < .30) with other items of child-mother
closeness. Item 3 and item 10 regarding child-father closeness did not correlate well with the scale overall either (r < .30) (see Appendix B for details). Also, a reliability analysis
regarding child-mother closeness showed that the reliability of the scale would be improved from Cronbach’s α = .86 to α = .88 if item 3 (reverse-coded) or item 10 was deleted. A reliability analysis concerning child-father closeness indicated that the reliability of the scale would be significantly improved from Cronbach’s α = .44 to α = .90 if item 10 was deleted. Therefore, item 3 and 10 were eliminated from the scale. The overall closeness score was the mean score of the remaining eight items combined (see Appendix B for details).
Regarding child-mother closeness, a principle axis factoring analysis with the remaining eight items indicated that all items form a single uni-dimensional scale: only one component had an eigenvalue over Kaiser’s criterion of 1 (EV = 4.83) and explained 60.35% of the variance. There was also a clear point of inflexion after this component in the scree plot. All items positively correlated with this component. The reliability of the scale was excellent with Cronbach’s α = .90 (M = 3.91, SD = .76). As to child-father closeness, all of the eight items were also loaded on one component that had an eigenvalue over Kaiser’s criterion of 1 (EV = 5.19) and explained 64.82% of the variance. The scree plot was clear with one point of inflexion after this component. The scale had high reliability with Cronbach’s α = .92 (M = 3.56, SD = .86). Therefore, it seems that the scale reliably measured participants’ child-mother and child-father relationship.
WeChat use. Based on the measure of Facebook use by Ledbetter et al. (2010), a total
of 10 items were created for Moments use and three items were devised for messaging use. Participants were firstly asked, “How often do you use WeChat to communicate with your mother/father?” If their answer had not been “never”, they were required to indicate how frequently they usually used each specific function to interact with their mother and father. All the items were measured on a 5-point scale ranging from 1 (never), 2 (rarely), 3 (occasionally), 4 (frequently) to 5 (very frequently).
Regarding WeChat communication between emerging adults and their mothers, a principle axis factoring analysis using direct oblimin indicated that 10 Moments items and three messaging items form a multidimensional scale: three components had an eigenvalue over Kaiser’s criterion of 1 and in combination explained 71.06% of the variance. The scree plot was slightly ambiguous showing two points of inflexion that would have justified retaining both components 2 and 5. Given the convergence of the scree plot and Kaiser's criterion on three components, three components were retained. Table 3 in Appendix B shows the factor loadings after rotation. The items that cluster on the same component suggested that component 1 represents direct communication with mothers on Moments, component 2 indirect communication with mothers on Moments, and component 3 interaction via
messaging. Component 1 had an excellent reliability with Cronbach’s α = .95. The reliability of component 2 was good with Cronbach’s α = .86. Component 3 had a good reliability with Cronbach’s α = .84. Component 2 measured the frequency of emerging adults’ status updates but whether their parents read the statuses were unknown, which could not represent
indicator of communication via Moments, and use component 3 to represent communication via messaging.
As to WeChat communication between emerging adults and their fathers, a principle axis factoring analysis using direct oblimin showed that all 13 items loaded on two
components that had an eigenvalue over Kaiser’s criterion of 1 and in combination explained 70.91% of the variance. There was a clear point of inflexion after this component in the scree plot. Table 4 in Appendix B shows the factor loadings after rotation. Unlike scales of
child-mother communication via WeChat, items regarding messaging use did not load on a separate component. Instead, it loaded on both component 1 representing direct
communication with fathers on Moments and component 2 indirect communication on Moments. In order to use same measurement scales for child-mother and child-father communication, it was decided to use the same items in WeChat communication between emerging adults and their parents to measure messaging use and Moments use in child-father communication.
Specifically, the three-item scale used to measure messaging use was, “I communicate with my mother/father via text messaging”, “I talk with my mother/father via voice
messaging”, and “I share pictures or videos with my mother/father via messaging”. Given that some participants did not use WeChat to contact mothers or/and fathers at all, missing values on relevant items were recoded to 1, which represented “never” in the 5-point scale. Then, The three items were averaged to form a scale. The reliability of the scale was excellent for
both child-mother communication with Cronbach’s α = .94 (M = 2.76, SD = 1.44) and child-father communication Cronbach’s α = .96 (M = 2.26, SD = 1.37)1.
A total of 8 items were used to measure Moments use (see Appendix A for details). For instance, “I reply my mother’s comments on my Moments” and “I 'like' my mother’s statuses” (see Appendix A for details). The overall Moments use score was an average score of all 8 items combined. The reliability of the scale was excellent for both child-mother communication with Cronbach’s α = .96 (M = 2.24, SD = 1.28) and child-father
communication with Cronbach’s α = .97 (M = 1.94, SD = 1.22)2.
Self-disclosure to parents. Participants’ self-disclosure to mothers and fathers was
separately measured using the slightly adapted 10-item Self-Disclosure Index (SDI) (Miller, Berg, & Archer, 1983). Items adapted form SDI had been successfully used in a study
focusing on Chinese university students (Liu & Brown, 2014). Participants were asked to rate their willingness to disclose the 10 types of intimate information to their mothers and fathers, for instance, “I disclose my personal habits to my mother” and “I disclose my deepest feelings to my mother” (see Appendix A for details). All items were assessed by a 5-point Likert scale (1 = strongly disagree to 5 = strongly agree).
Regarding participants’ level of disclosure to mothers, a principle axis factoring analysis with 10 items indicated that all items load on a single dimension: only one
component had an eigenvalue over Kaiser’s criterion of 1 (EV = 7.20) and explained 71.95%
1 The reliability of the messaging scale was improved after recoding missing variables, from Cronbach’s α = .84 to α =
.94 for child-mother communication, and from Cronbach’s α = .89 to α = .96 for child-father communication.
2 The reliability of the Moments scale was slightly improved after recoding missing variables, from Cronbach’s α = .95
of the variance. Also, a clear point of inflexion after this component was shown in the scree plot. All items positively loaded on this component (loadings ≥ .79). The reliability of the scale was perfect with Cronbach’s α = .96 (M = 3.56, SD = .92).
As to participants’ level of disclosure to fathers, all of the 10 items were also loaded on one component that had an eigenvalue over Kaiser’s criterion of 1 (EV = 7.64) and explained 76.44% of the variance. In the scree plot, there was a clear point of inflexion after this component. The scale had high reliability with Cronbach’s α = .97 (M = 3.24, SD = .96). All items positively correlated with this component (loadings ≥ .59).
Geographic distance to parents. One question was used to measure child-parent
living distance, “How close to your biological mother/father do you live? If you live in more than one residence, (e.g., university dormitory and parental home), it refers to where you mainly live” in the beginning of the survey. The answer categories ranged from 1 (same house), 2 (same city), 3 (same province), 4 (same country) to 5 (further away). This scale was adapted from a survey regarding communication and closeness (Lee, 2002). Because the target group of the study was the emerging adults who did not live together with their parents, participants who chose “same house” were removed from the data. In the analysis, “same city” and “same province” were merged into one category due to small sample sizes (N < 25). Geographic distance was then recoded in to 1 = same province (mother 23.1% and father 22.1%), 2 = same country (mother 40.0% and father 40.5%), 3 = different countries (mother 36.9% and father 37.4%). The mean of child-mother distance score was 2.14 (SD = .76) and child-father living distance was 2.15 (SD = .76).
Communication via other channels (covariate). Participants’ communication with
their parents through alternative communication channels (i.e., face-to-face communication, voice call, voice messaging, video chats, text messaging, interaction on social media) was measured on 6-point scales ranging from 1 (never) to 6 (multiple times a day). Given that most of Chinese university students contacted their parents once a week (He & You, 2006), the measurement scale from the study of Lee (2002) was most applicable to the current study. An option “2-3 times per week” was added between “daily” and “weekly”, and the original description “less often” was replaced by “less than once a month” in order to allow for an adequate distribution of the responses.
Communication by voice messaging highly correlated with the variable communication via WeChat messaging (see Appendix C for details; child-mother
communication, r = .75, p < 0.01, and child-father communication, r = .72, p < 0.01). This may be because WeChat voice messaging was the main and even the only voice messaging used by participants as WeChat first introduced the ability to ‘send voice messages’ in the smartphone application (Wang & McDonald, 2013). Therefore, it was decided to remove voice messaging from the overall communication via other channels to the correlation between WeChat use and overall communication, which would have resulted in substantial collinearity. Thus, the score of overall communication via other channels was an average score of the remaining five channels. Regarding child-mother overall communication via other channels, the scale resulted in a Cronbach’s α of .67 (M = 3.60, SD = 1.28). The scale of child-father overall communication via other channels had a good reliability with Cronbach’s α = .77 (M = 3.00, SD = 1.34).
Control variables. Participants' age was measured by an open-ended question, "what
is your age". Dichotomous variables were adopted to determine participants’ gender (0 = male and 1 = female) and their parents’ marital status (0 = separate and 1 = together). Moreover, participants’ marital status was assessed by a nominal variable with three categories (0 = single, 1 = married and 2 = divorced). Family standard of living was measured by a question adopted from the World Values Survey (WVS) on a 4-point scale (1 = save money, 2 = just get by, 3 = spend some saving, and 4 = spend savings and borrow money).
Results Zero-Order Correlations Between Main Variables
Bivariate zero-order correlations of all constructs and covariates are illustrated in Table 7 in Appendix D. Below, important correlations were presented in detail.
As shown in Table 7, child-mother closeness had significant, moderate positive associations with communication with mothers via WeChat messaging (r = .39, p < .01), Moments (r = .34, p < .01), other channels (r = .41, p < .01), voice calls (r = .32, p < .01), and video chats (r = .30, p < .01). Child-mother closeness also weakly and positively correlated to face-to-face communication with mothers (r = .14, p < .05). This means that the closer
participants feel to their mothers, the more they communicate by WeChat messaging, Moments, voice calls, video chats and in person, and vice versa.
Child-father closeness had significant moderate positive associations with child-father communication via WeChat messaging (r = .43, p < .01), Moments (r = .40, p < .01), voice calls (r = .34, p < .01), video chats (r = .36, p < .01), as well as other channels (r = .49, p < .01). Besides, it had small and positive correlations with face-to-face communication with
fathers (r = .17, p < .01) and parents’ marriage status (r = .27, p < .01). That is to say, the closer participants feel to their fathers, the more they communicate by WeChat messaging, Moments, voice call and in person, and the other way around. Participants whose biological parents are together have a closer relationship with their fathers than those whose parents are separated. Yet, child-father closeness was weakly and negatively related to family’s standard of living (r = -.13, p < .05). It indicates that participants from less wealthy families felt closer to their fathers than those from richer families.
Moreover, participants’ closeness to mothers moderately and positively correlated to closeness to fathers (r = .48, p < .01). There was also a significant, strong association between child-mother and child-father communication via WeChat messaging (r = .69, p < .01) as well as Moments (r = .75, p < .01). Additionally, participants’ age was significantly weakly related to child-mother and child-father closeness, as well as child-mother communication via
WeChat messaging (r < .20, p < .05). Participants’ gender had significant small correlations with child-mother communication through WeChat messaging (r = .20, p < .01) and Moments (r = .15, p < .05).
It is worth noting that communication via WeChat messaging was significantly, very strongly related to communication through Moments, for both child-mother (r = .80, p < .01) and child-father communication (r = .84, p < .01). Therefore, the study investigated
child-parent closeness’s association with WeChat messaging and Moments in two separate models to avoid collinearity in further analysis.
Model Specification
All hypotheses were tested using Structural Equation Modelling with AMOS 21 and maximum likelihood estimation. Model fit was assessed by inspecting chi-square test, the comparative fit index (CFI), and the root mean square error of approximation (RMSEA). An acceptable model fit is indicated by a non-significant chi-square value, a CFI larger than .90, and a RMSEA smaller than .08 (Byrne, 2010). It must be noted that the chi-square test often significant even when misspecifications are quite small (Byrne, 2010).
Following from the hypotheses, there exists a mutual interaction between chid-parent closeness and communication via WeChat including messaging and Moments: the closer the child-parent relationship that emerging adults feel, the more emerging adults use WeChat messaging and Moments to communicate with their parents, and vice versa. Thus, two non-recursive path models were estimated, one for WeChat messaging use (see Figure 2 in Appendix E) and the other for Moments use (see Figure 3 in Appendix E). In both models, communication via other channels, participants’ gender, parents’ marriage status, and
family’s standard of living were entered as control variables (i.e., exogenous) variables. In the model regarding messaging use, communication with mothers via messaging, communication with fathers via messaging, child-mother closeness and child-father closeness were entered as endogenous variables. In the model concerning Moments use, communication with mothers via Moments, communication with fathers via Moments, child-mother closeness and
child-father closeness were entered as endogenous variables.
The hypothesized model concerning WeChat messaging use was tested first. The model fit well, χ2 (2) = .852, p = .653, CFI = 1.00, RMSEA = .00, 90% CI [.00, .11]. However,
the model resulted in an improper solution with standardized regression coefficients bigger than one. The b* of the path from communication with mothers via WeChat messaging to child-mother closeness was 1.17. Also, the b* of the path from communication with fathers via WeChat messaging to child-father closeness was 1.04.
The overall fit of the model regarding Moments use was acceptable, with χ2 (2) = 3.458, p = .177, CFI = 1.00, RMSEA = .06, 90% CI [.00, .17]. Yet, the model led to an improper solution with several standardized regression coefficients bigger than one, such as the path from communication with mothers on Moments to child-mother closeness (b* = 2.46, p = .028).
As shown in Table 5 and Table 6, it appears that the improper solution with a standardized regression coefficient bigger than one stems from high collinearity between communication via other channels and communication via WeChat. Thus, an alternative model was estimated in which each communication channel was entered as a separate variable instead of using a summary of the five channels. This was done to identify and remove those communication channels which produced the collinearity and the improper solution. Face-to-face communication and interaction via voice calls had small associations with both WeChat messaging and Moments (r < .30, p < .05). The model including
face-to-face communication and interaction via voice calls was treated as a base model. Then, the three other communication channels were added to the base model individually to test whether the new models work.
The results showed that the model regarding WeChat Messaging use, which included face-to-face communication, interaction via voice calls and video chats as covariates, had an
acceptable model fit with χ2 (12) = 24.31, p = .018, CFI = .99, RMSEA = .07, 90% CI [.03, .11]. Also, model with Moments use fit well, with χ2 (12) = 20.74, p = .054, CFI = .99, RMSEA = .06, 90% CI [.00, .11]. More importantly, the two modified models did not include any improper solutions. Technically, the three remaining channels do not overlap with the two functions of WeChat involved in the study, namely messaging and Moments. Consequently, it was decided to use the models, which included face-to-face communication, interaction via voice calls and video chats as covariates, as final models in further analysis.
Testing the Relationship between WeChat Use and Child-Parent Closeness
The final models are depicted in Figure 4 and Figure 5. Main correlations among covariates and paths from exdogenous variables to endogenous variables are shown in Appendix F.
Figure 4. Non-recursive path model regarding WeChat messaging use and child-parent closeness, with video chats, face-to-face communication and voice call as covariates. Coefficients represent standardized betas. Solid lines represent significant effect at least p < .05. Dash lines represent effects are not significant. M stands for mother and D for fathers.
Figure 5. Non-recursive path model regarding WeChat Moments use and child-parent closeness, with video chats, face-to-face communication and voice call as covariates. Coefficients represent standardized betas. Solid lines represent significant effect at least p < .05. Dash lines represent effects are not significant. M stands for mother and D for fathers.
First, the hypothesized model concerning WeChat messaging use in Figure 4 was tested. This model was over-identified with dfM = 12 and had an acceptable model fit, with χ2
(12) = 24.31, p = .018, CFI = .99, RMSEA = .07, 90% CI [.03, .11]. Thus, no modification to the model was necessary.
Hypothesis 1a, which predicted that child-mother closeness would positively predict emerging adults’ communication with mothers via WeChat messaging, was rejected (b* = -.28, p = .179). The hypothesis 2a, which indicated that child-father closeness would be positively related to emerging adults’ communication with fathers via WeChat messaging, was also rejected (b* = -.32, p = .174).
Hypothesis 3a was supported. As Figure 4 shows, emerging adults’ communication with mothers via WeChat messaging positively predicted the closeness that emerging adults
felt towards their mothers (b* = .55, p = .009). The strength of this effect was strong. Besides WeChat messaging, interaction through voice call also weakly affected child-mother
closeness (b* = .28, p < .001). Likewise, hypothesis 4a, which stated that the more emerging adults contacted fathers via WeChat messaging the closer relationship they reported, was supported (b* = .54, p = .004). The strength of this effect was strong. Child-father closeness was also slightly influenced by interaction via voice calls (b* = .18, p = .008).
By comparing effect size, it shows that the effect of WeChat messaging use on
child-mother closeness (b* = .55) was almost equal to the effect on child-father closeness (b* = .54). Accordingly, a constrained model indicated that there was no significant difference among child-mother and child-father closeness, with χ2DIFF(1) = 1.10, p = .294. Hypothesis
11a, which predicted the positive effect of WeChat messaging on child-parent closeness was stronger for child-father relationship than child-mother relationship, was thus rejected.
Second, the hypothesized model regarding WeChat Moments use in Figure 5 was investigated. The fit of the model was good, with χ2 (12) = 20.74, p = .054, CFI = .99, RMSEA = .06, 90% CI [.00, .11].
Hypothesis 1b, which stated that child-mother closeness would positively predict emerging adults’ communication with mothers on Moments, was rejected (b* = -.02, p = .886). Hypothesis 2b, which indicated that child-father closeness would be positively related to emerging adults’ communication with fathers on Moments, was rejected as well (b* = -.10, p = .488).
Moreover, hypothesis 3b, which predicted that child-mother communication on Moments positively influenced their closeness, was rejected. The effect from Moments
communication on the child-mother closeness was not significant (b* = .23, p = .135). Only interaction via voice calls had a significantly small effect on the closeness (b* = .27, p < .001). Yet, hypothesis 4b, which predicted that child-father communication on Moments positively influenced their closeness, was supported (b* = .31, p = .029). Thus,
communication with fathers on Moments positively, moderately predicted child-father
closeness. Besides, interaction via voice calls had a small effect on the closeness as well (b* = .21, p = .001).
It seems that the effect of Moments use on child-parent closeness was stronger for child-father relationship (b* = .31) than child-mother relationship (b* = .23) as hypothesis 11b predicted. But, this difference is not significant χ2DIFF(1) = 2.66, p = .103. Therefore,
hypothesis 11b was rejected.
As shown in Figure 4 and Figure 5, communication via WeChat messaging had greater effects on both child-mother closeness and child-father closeness compared to
communication via Moments. The result was plausible considering that model structures, the sample, measures on all the variables except for messaging and Moments use were the same. Thus, hypothesis 5 and hypothesis 6, which assumed that communication with parents via WeChat messaging had a greater positive effect on child-parent closeness than that via Moments, were supported.
Testing the Moderating Effects of Self-Disclosure Tendencies
Table 12 in Appendix G shows specific effect sizes of WeChat communication on child-parent closeness as per self-disclosure tendency.
Hypothesis 7a predicted that the positive effect of WeChat messaging on child-mother closeness was stronger for emerging adults with higher self-disclosure to mothers than those with lower self-disclosure to mothers. In order to investigate the hypothesis, self-disclosure to mothers was dichotomized by means of a median split. Then a two-group analysis was
conducted with self-disclosure to mothers as the grouping variable. The fit of the
unconstrained model for the two groups was acceptable, with χ2 (24) = 41.28, p = .016, CFI = .98, RMSEA = .06, 90% CI [.03, .09]. The effect of WeChat messaging on child-mother closeness was moderate for the low self-disclosure group, but weak for the high
self-disclosure group. Then the paths from messaging use to child-mother closeness were constrained to be equal between two groups. The result showed that the paths between the two groups were not significantly different, χ2DIFF(1) = 1.38, p = .240. It means
self-disclosure does not moderate the effect of communication through messaging on child-mother closeness, so hypothesis 7a was rejected.
Hypothesis 8a anticipated that the positive effect of WeChat messaging on child-father closeness was stronger for emerging adults with higher self-disclosure tendencies than those with lower self-disclosure tendencies. Again, a nested model comparison was conducted. The fit of the unconstrained model for the two groups (high self-disclosure to fathers vs. low self-disclosure to fathers) was acceptable, with χ2 (24) = 50.05, p = .001, CFI = .96, RMSEA = .07, 90% CI [.04, .10]. There was a strong effect of WeChat messaging on closeness for the low self-disclosure group but a weak effect for the high self-disclosure group. Yet, results of the model with cross-group constraints indicated there was no significant difference between
high and low self-disclosure groups, with χ2DIFF(1) = 2.63, p = .105. Hypothesis 8a was thus
rejected.
Hypothesis 7b predicted that the positive effect of Moments on child-mother closeness was stronger for emerging adults with higher levels of self-disclosure to mothers than those with lower levels of self-disclosure tendencies. A nested model comparison was conducted. The unconstrained model for the high and low self-disclosure groups had an acceptable model fit, with χ2 (24) = 49.54, p = .002, CFI = .97, RMSEA = .07, 90% CI [.04, .10]. The assumed effect was weak for participants with low self-disclosure, and very weak for those with high self-disclosure. However, the difference between low and high self-disclosure groups was not significant, with χ2DIFF(1) = .39, p = .530. Hypothesis 7b was thus rejected.
Hypothesis 8b stated that the positive effect of Moments on child-father closeness was stronger for emerging adults with higher self-disclosure to fathers than those with lower self-disclosure tendencies. Again, a nested model comparison was conducted. The results showed that the fit of the unconstrained model was good, with χ2 (24) = 34.83, p = .071, CFI = .99, RMSEA = .05, 90% CI [.00, .08]. The effect of Moments on child-father closeness was weak for participants with high self-disclosure and very weak for those with low
self-disclosure. However, there was no significant difference between high and low self-disclosure groups, with χ2DIFF (1) = 0.00, p = 1.00. Hypothesis 3db was rejected too.
Testing the Moderating Effects of Geographic Distance
Table 13 in Appendix G shows specific effect sizes of WeChat communication on child-parent closeness as per geographic distance.
Hypothesis 9a predicted that the positive effect of WeChat messaging on child-mother closeness was stronger for emerging adults living further away from mothers than those living close to their mothers. In order to investigate the hypothesis, a three-group analysis was conducted with geographic distance to mothers as the grouping variable. A total of 45 participants lived in the same province with their mothers, 78 lived in the same country but different provinces, and 72 lived in different countries. The fit of the unconstrained model was good, with χ2M(36) = 54.91, p = .023, CFI = .98, RMSEA = .05, 90% CI [.02, .08]. There
were weak effects of WeChat messaging on child-mother closeness for participants living in the same province and those living in the same country with their mothers. But the assumed effect was extremely strong for participants who lived in different countries with their mothers.3 Then the paths from messaging use to child-mother closeness were constrained to be equal among three groups, but the result showed no significant difference among the three paths, with χ2DIFF (2) = 3.83, p = .148. It indicates that the effects of geographic distance to
mothers on child-mother closeness were not significantly different between participants living far away and those living close to mothers. Thus, hypothesis 9a was rejected.
Hypothesis 10a predicted that the positive effect of WeChat messaging on child-father closeness was stronger for emerging adults living further away from fathers than those living close to their fathers. Again, a nested model comparison was carried out. There were 43 participants living in the same province with their fathers, 79 living in the same country but different provinces, and 73 living in different countries. The fit of the unconstrained model
3 Models regarding messaging use led to an improper solution with a few standardized regression coefficients close
was acceptable, with χ2 (36) = 57.97, p = .022, CFI = .98, RMSEA = .05, 90% CI [.02, .08]. The effect of WeChat messaging on child-father closeness was strong for participants living in the same province and those living in different countries with fathers. Communication via messaging hardly affected the closeness for participants living in the same country but different provinces with fathers. Yet, there was no significant difference on the assumed effect on child-father closeness among the three groups, with χ2DIFF (2) = 2.82, p = .244.
Consequently, Hypothesis 10a was rejected.
Hypothesis 9b predicted that the positive effect of Moments on child-mother closeness was stronger for emerging adults living further away from mothers than those living close to their mothers. A nested model comparison was conducted. The unconstrained model had an acceptable fit, with χ2 (36) = 57.56, p = .013, CFI = .98, RMSEA = .06, 90% CI [.03, .08]. The effect of Moments on child-mother closeness was strong for participants living in the same country, and moderate for participants living in different countries as fathers. But, it had no effect for participants living in the same province with fathers. Results of the model with cross-group constraints indicated the difference among the three geographic distance groups was not significant, with χ2DIFF(2) = 1.62, p = .446. Therefore, hypothesis 9b was rejected.
Hypothesis 10b predicted that the positive effect of Moments on child-father closeness was stronger for emerging adults living further away from fathers than those living close to their fathers. The results showed that the fit of the unconstrained model was acceptable, with χ2
(36) = 58.54, p = .010, CFI = .97, RMSEA = .06, 90% CI [.03, .08]. As shown in Table 13, the assumed effect was strong for participants living in different countries, moderate for participants living in the same country and those living in the same province as fathers.
However, there was no significant difference on the effect of Moments on child-father closeness among participants living in the same province, participants living in the same country, and participants living in different countries with fathers, χ2DIFF (2) = .67, p = .716.
Hypothesis 10b was thus rejected.
Discussion and Conclusion
This study was the first to examine the interaction between Chinese child-parent closeness and their communication via WeChat messaging and Moments. Findings rejected the expectation that child-mother closeness positively affects how frequently emerging adults communicate with their mothers via WeChat messaging and Moments (H1). The study also found little support for the positive effect of child-father closeness on their communication via WeChat messaging and Moments (H2). However, the results confirmed the expectation that communication via WeChat messaging has a positive effect on the closeness emerging adults feel to their mothers (H3a) and fathers (H4a). In other words, while more frequent communication via WeChat messaging increases the closeness emerging adults feel to their parents, the emotional closeness does not have an effect on how frequently they interact via WeChat messaging. Therefore, the results partially supported Homan’s system of
interpersonal relations (1950), which stated that the closer the relationship between emerging adults and their parents, the greater their need to interact with one another. In turn, the more frequently emerging adults communicate with their parents, the closer they feel towards each other. It appears that although several previous studies reported positive effects of
child-parent closeness on their contact (e.g., Lawton, 2014), the first part of Homans’ system of interpersonal relations may, in its generality, no longer be plausible to explain the effects of
child-parent closeness on communication via WeChat. Yet, the second part of the Homans’ theory was confirmed in the present study, which predicted positive effects of emotional closeness on communication. Also, these results were consistent with many previous studies focusing on other forms of communication, such as call phones (Gentzler et al., 2011) and messaging (Markopoulos et al., 2004).
Although several early studies failed to find an influence of messaging communication on interpersonal relationships (e.g., Jin & Pena, 2010), this study found a strong positive effect of communication via WeChat text and voice messaging on child-mother and
child-father closeness. This may be because WeChat text and voice messaging makes up for two disadvantages of smartphone text messaging: (1) limited amount of information carried by one text message and (2) voice is not available during the communication (Jin & Pena, 2010). Specifically, the WeChat “hold to talk” voice messaging feature enables users to easily send a voice-based message that can transmit much more information than a text message. Also, voice and intonation help strengthen expression and understanding. (Daft & Lengel, 1986; Wang & McDonald, 2013).
In addition, the positive effect of communication with fathers via Moments on
child-father closeness was confirmed (H4b), while the effect of communication with mothers via Moments on child-mother closeness was rejected (H3b). The main function of Moments is sharing individuals’ life streams, feelings, thoughts or experiences with others. The
information revealed on Moments is always superficial information because it is difficult to elaborate something in a few sentences and pictures, as well as to disclose important
mothers and tell them what is happening in their life (Yang, 2014), this superficial
information of the emerging adults is accessible to mothers in daily communication. So, the superficial interaction on Moments hardly enhances the child-mother closeness. However, most emerging adults hardly take initiative to contact fathers and have limited common topics to talk about (Yang, 2014). Hence, the superficial communication via Moments still plays an important role in the child-father closeness. More exact reasons should be investigated in future study.
Consistent with media richness theory (Daft & Lengel, 1986), the study demonstrated that communication with parents via WeChat messaging has a greater positive effect on child-parent closeness than that via Moments, irrespective of parents’ gender (H5 and H6). It appears that text and voice messaging is more efficient in increasing child-parent closeness than the social networking service, Moments, due to the higher ability of reproducing cues (e.g., voice and intonation).
Besides, the study looked at the moderating roles of emerging adults’ self-disclosure to their parents, the child-parent geographic distance, and the gender of parents. Little support was found for the moderating effect of emerging adults’ self-disclosure towards their mothers (H7) and fathers (H8). There was no significant difference on the effect of WeChat on
child-parent closeness between low and high self-disclosure groups. It seems that the trait-like self-disclosure is not important in the effect of WeChat communication on child-parent
closeness. Interestingly, it was found that the effect of messaging on child-parent closeness and the effect of Moments on child-mother closeness were stronger for low self-disclosure groups than the high self-disclosure groups. It is contrary to the hypotheses and against the
idea that individuals’ willingness and ability to reveal themselves to others during communication helps the development and the maintenance of close relationships (e.g., Altman & Taylor, 197; Greene et al., 2006). The stronger influence of WeChat
communication on child-parent closeness for individuals with less self-disclosure tendencies may be because they feel more comfortable and gain more benefits from the asynchronous (as opposed to real-time) communication compared to individuals with higher self-disclosure tendencies (Chan, 2011). Specifically, individuals with low self-disclosure are less likely to express themselves during the interpersonal communication. Compared to real time
communication (e.g., face-to-face interactions), asynchronous communication on Moments reduces demands of spontaneity and then potentially makes them feel more relaxed.
Findings rejected the assumption that the positive effect of WeChat use on
child-parent closeness is stronger for emerging adults who live further away from parents than those who live closer to their parents (H9 and H10). No significant differences were found among participants who live in the same province as parents, participants who live in the same country but different provinces as parents, and participants who live in different countries as parents. The study argued that there were increasing effects of WeChat
communication on child-parent closeness when they made less visits and phone calls due to the growing time and economic cost of a long-distance relationship (Lee, 2002). However, the study might ignore the fact that a few other forms of online communication besides WeChat are also widely used in long-distance familial relationships, and compensate for the lack of face-to-face communication and phone calls, such as Tecent QQ (a major Chinese
Internet-based instant messaging platform) and Sina weibo (Shen, 2012, 2014). This communication may dilute the effect of WeChat communication on child-parent closeness.
Moreover, the results did not find significant difference on the effects of WeChat on child-parent closeness between child-father and child-mother relationships (H11). The effect of WeChat messaging use on child-mother closeness was almost equal to the effect on child-father closeness, while the effect of Moments was stronger for child-father relationship than child-mother relationship. This may be because WeChat messaging has become one of the basic communication channels in both child-mother and child-father communication while Moments is a supplementary communication channel that is not widely used. Besides the basic channels, fathers hardly communicate with their children in emerging adulthood via extra channels. Yet, mothers are more active in interacting with their children via different communication channels (Sawhney & Gomez, 2000), so the effect of communication via Moments on child-mother closeness is diluted and then smaller than the effect on child-father closeness.
Limitations
Three limitations of the present study must be addressed. The first limitation is related to the measure of communication via other channels in the current study. Only two items of communication via other channels (i.e., face-to-face communication and interaction through voice calls) were included in the models as covariates, which influences the meaning of the results and the corroboration of the hypotheses. Theoretically, all the main communication forms used by emerging adults and their parents should be controlled since they may influence child-parent closeness as well.
Second, the assumptions in the models were tested with cross-sectional data in which all the variables were collected more or less simultaneously. Therefore, it cannot be
ascertained whether the demonstrated effects of communication via WeChat messaging on child-parent closeness denote a causal relationship (Bryman, 2008). Thus, further studies should use a longitudinal design to examine the causality between communication via WeChat messaging and child-parent closeness.
Third, this study was based on data from a non-probability convenience sample. Consequently, the sample may over-represent participants with backgrounds similar to the researcher and her acquaintances. This assumption is substantiated by the high proportion of participants (22.1%) who come from Sichuan province, and well-educated participants (98.5%) with at least a Bachelor’s degree. Therefore, the findings cannot be generalised beyond the given research context (Bryman, 2008).
Implications
Overall, the present study provides both theoretical and practical contributions to the fields of communication study and family study. First, it fills the gap in the academic research on the effect of WeChat, as a newly popular communication technology, on child-parent relationships. Second, it provides implications for further research on communication and interpersonal relationships. Third, it may inspire parents with ways of communicating with their children in emerging adulthood. Specifically, communicating in a young person’s way can be a good choice, such as writing “statuses” and sharing information with their children on social media.
