The role of distance in online social networks: A case study of urban residents in Nanjing, China

The role of distance in online social networks: A case study of urban residents in Nanjing, China

Cities xxx (xxxx) xxx–xxx Contents lists available at ScienceDirect Cities journal homepage: www.elsevier.com/locate/cities The role of distance in...

249KB Sizes 0 Downloads 37 Views

Cities xxx (xxxx) xxx–xxx

Contents lists available at ScienceDirect

Cities journal homepage: www.elsevier.com/locate/cities

The role of distance in online social networks: A case study of urban residents in Nanjing, China ⁎

Bo Wanga, Feng Zhenb, , Becky P.Y. Looa,c a

Department of Geography, The University of Hong Kong, China School of Architecture and Urban Planning, Nanjing University, China c HKU-Shenzhen Institute of Research and Innovation (HKU-SIRI), China b

A R T I C L E I N F O

A B S T R A C T

Keywords: Information and communication technologies Social network Urban transformation, China

Highly advanced information and communication technologies have reshaped the common ways urban residents interact with each other. With the widespread use of online social networking websites, research interest in the evolving spatial concepts (such as distance) in new digital age has grown exponentially. Data collected by Sina microbloggers from adult residents in Nanjing, China reveal that urban residents not only are more likely to build relationships with local and acquaintance users but also to interact with them more frequently. In other words, spatial and relational distances that play important roles in traditional Chinese social networks also exist in contemporary online social networks. Furthermore, our regression analysis reveals how the roles of spatial and relational distances in the online social networks of urban residents relate to the context of urban transformation in contemporary China. The findings contribute to a more in-depth understanding of the effects of informatisation and urban transformation on the social networks of urban residents.

1. Introduction Engaging in interpersonal interactions, whether in the present, the past or the future, is fundamental to human life, because although not everyone has a full-time job, every human being lives in a social network (Loo, 2012; Sgroi, 2008). Cities, being ‘melting pots’, have enjoyed the advantage of providing urban residents platforms to build multiple social networks and thereby develop interaction opportunities (Glaeser, 2011). Spatial and relational distances1 have long been known to heavily influence the formation and development of urban residents' social networks before the Internet (Mok & Wellman, 2007; Sgroi, 2008; Wellman & Leighton, 1979). However, the key roles of spatial and relational distances in social networks have been increasingly challenged by the rapid development of information and communication technologies (ICTs), which have attracted substantial attention from urban scholars in the West (e.g., Cairncross, 2001; Mok, Wellman, & Carrasco, 2010; Thulin & Vilhelmson, 2005). Nowadays, there are many electronic communication modes, including (mobile) phone calls, instant message services, emails and, more recently, popular online social networking websites (OSNs). Given their low cost, these electronic communication modes are expected to

influence the relationship between spatial and relational distances and people's social networks (Cairncross, 2001). Despite the large and growing evidence of the role of spatial and relational distances in telephone and email communications (e.g., Carrasco, Miller, & Wellman, 2008; Mok et al., 2010; Mok & Wellman, 2007; Thulin & Vilhelmson, 2005; Tillema, Dijst, & Schwanen, 2010), our knowledge of interpersonal interactions in OSNs (such as Twitter, Facebook and LiveJournal in Western countries and Weibo in China) remains inadequate, especially considering their global popularity (Huang & Sun, 2014; Takhteyev, Gruzd, & Wellman, 2012; Zhen, Wang, & Chen, 2016). Unlike telephone and email services, OSNs are regarded as cheap, participative, interactive, open and transparent (Dekker, Engbersen, & Faber, 2016; Huang & Sun, 2014; Kaplan & Haenlein, 2010); they present convenient ways to interact with others both locally and globally. Moreover, OSNs have provided new opportunities for maintaining and extending interpersonal interactions among both acquaintances and strangers2 more easily. However, the ties formed through OSNs appear much weaker than those formed in the real world (Huang & Sun, 2014; Takhteyev et al., 2012) and may therefore shape the ways users interact with each other differently. Hence, it is reasonable to question the roles of spatial and relational distances in interpersonal



Corresponding author at: Room 618, Wenke Building, School of Architecture and Urban Planning, Nanjing University, Nanjing, China. E-mail address: [email protected] (F. Zhen). 1 Relational distance refers to the difference in the somewhat close and loose ties for different relationships, such as immediate/extended kin, friends, neighbours and strangers. In this study, we examine the relational distance between the relationship of acquaintances and strangers in the real world. 2 In Twitter and Weibo, a user can choose to follow another without the latter's permission. https://doi.org/10.1016/j.cities.2018.02.020 Received 9 August 2017; Received in revised form 10 January 2018; Accepted 24 February 2018 0264-2751/ © 2018 Published by Elsevier Ltd.

Please cite this article as: Wang, B., Cities (2018), https://doi.org/10.1016/j.cities.2018.02.020

Cities xxx (xxxx) xxx–xxx

B. Wang et al.

and the frequency of e-mail communication also gradually drops over longer distances (Mok et al., 2010). A similar phenomenon has also been identified in the Netherlands (Tillema et al., 2010) and Sweden (Thulin & Vilhelmson, 2005). The friction sensitivity of telephone and email correspondence to spatial distance could be explained by the somewhat strong social ties in these communication modes. Generally, people prefer to store the phone numbers of their acquaintances, and people who have more face-to-face interactions are more likely to make telephone calls (Thulin & Vilhelmson, 2005; Tillema et al., 2010). Some studies have also explored the effects of relational distance on email communications. A study on urban youth in Sweden has found that people tend to communicate via email more with members they already know in real life (Thulin & Vilhelmson, 2005). For OSNs, the relationship between users tend to be much weaker than that among people making telephone calls or having regular email communications (Dekker et al., 2016; Huang & Sun, 2014; Kaplan & Haenlein, 2010). Hence, the role of distance can also be very different in the virtual social networks of OSNs than other online social networks or in the real world. As Thulin and Vilhelmson (2005) suggest, OSNs offer numerous opportunities for people to have new cyberfriends they have never met face-to-face. Besides, OSNs, as a media-rich way of communications, present important platforms for users to share and exchange information regarding specialised interests in areas such as sports, arts, politics and economy (Dekker et al., 2016; Zhen et al., 2016). However, these online relationships formed in OSNs may remain entirely virtual, with no physical interactions among users in the real world (Thulin & Vilhelmson, 2005). Yet, it is also true that the online relationships of OSNs users may come directly from their social networks in the real world, with a group of users located in the same physical space and sharing the same social network (Kellerman, 2016; Loo, 2012; Sharmeen et al., 2014). Using the geographical location information in user profiles, some studies suggest that OSNs users (e.g., those on Twitter and LiveJournal) tend to build more relationships with other users within a shorter spatial distance in the real world, indicating that spatial distance still matters (e.g., Liben-Nowell, Novak, Kumar, Raghavan, & Tomkins, 2005; Takhteyev et al., 2012). As geographers, we understand that the evolving role of ICTs and their impacts on the interpersonal interactions of OSNs users may not be the same in different contexts (Thulin & Vilhelmson, 2005; Wang et al., 2013). Hence, whether findings based on a Western context can be generalised to China should be verified, as the latter has experienced rapid informatisation and urbanisation simultaneously in the past two decades (Loo & Wang, 2017a). In addition, there are many research gaps, including the role played by spatial distance among residents of different segments. Previous studies have suggested that the differences in online activity experience of individuals with different socioeconomic characteristics could be substantial (Kellerman, 2016; Loo, 2012). Furthermore, urban residents in the same city may use OSNs differently (Thulin & Vilhelmson, 2005; Tillema et al., 2010). With more time spent online, these people may reduce the time spent with others physically (Huang & Sun, 2014; Thulin & Vilhelmson, 2005). In this case, urban residents who spent more time in OSNs may have less interactions with locals and acquaintances. Besides, few studies have examined relational distance in OSNs (Laniado, Volkovich, Scellato, Mascolo, & Kaltenbrunner, 2017; Wang et al., 2013). Based on this understanding, the following hypothesis is developed:

interactions in these virtual social networks. Urban scholars have devoted considerable attention to investigating the effects of urban transformation on social relations in the real world (e.g., Fischer, 1982; Forrest & Yip, 2007; Hofferth & Iceland, 1998; Kearns & Forrest, 2000; Zhang, Wu, Zhong, Zeng, & Wang, 2017). Various studies point out that residential environment have a substantial impact on the daily social interactions of urban residents (Amin & Thrift, 2002; Fischer, 1982; Sharmeen, Arentze, & Timmermans, 2014; Wang, Zhang, & Wu, 2016; Whyte & Parish, 1985). However, very few studies have focused on the impacts of urban transformation on the interpersonal interactions of urban residents in virtual networks such as OSNs. The limited research studies on this topic were based on the developed countries where informatisation followed urbanisation. As a developing country undergoing rapid informatisation and urbanisation simultaneously (Loo & Wang, 2017a; Wu, 2015), China represents a natural ‘laboratory’ ideal for studying the impacts of urban transformation on the OSNs of urban residents. Specifically, this study examines how the roles of spatial and relational distances in urban residents' OSNs relate to the context of urban transformation in contemporary China. It contributes to a better understanding of the effects of ICTs on social networks in the broader context of urban transformation. Given the widespread adoption of OSNs by urban residents, an empirical study on the interpersonal interactions of urban residents in OSNs should also offer insights into the effects of informatisation and urban transformation on the social networks of urban residents. This study utilises data on the interpersonal interactions of adult Nanjing residents on the Sina Weibo platform, the most popular OSNs in China, to explore the roles of spatial and relational distances in OSNs. In particular, we examine the virtual ties between the urban residents and local/non-local users and acquaintance/stranger users on Sina Weibo. The differences among the relationships/interaction frequencies between locals and non-local users illustrate the role of spatial distance, while the differences between acquaintance and stranger users illustrate the role of relational distance. The following research questions are considered in this study: (a) Do spatial and relational distances matter in the online social networks of urban residents? If so, how? (b) How do socioeconomic conditions and levels of Internet use experience influence the roles of spatial and relational distances in the virtual interpersonal interactions of urban residents? Finally, (c) how have the roles of spatial and relational distances in urban residents' interpersonal interactions been mediated in the context of urban transformation? 2. Literature review Social contacts have benefited from spatial proximity because frequent face-to-face contacts among spatially dispersed ties were naturally hindered by spatial distance. Urban residents traditionally have more face-to-face communications with locals and acquaintances than with non-locals and strangers (Glaeser, 2011). Advancements in transportation have promoted interpersonal interactions across cities and increased opportunities to communicate with strangers; however, these communications were limited by the speed and cost of travel (Mok et al., 2010). A 1978 survey conducted in Toronto identified a marked drop in the frequency of face-to-face contacts at about 5 miles and a steady decrease at greater distances (Mok & Wellman, 2007). The advancement of ICTs has triggered a heated debate over the impacts of new technology on the social networks of urban residents in the real world (Dekker et al., 2016; Huang & Sun, 2014; Mok et al., 2010; Sgroi, 2008). Despite the proclamations of ‘the death of distance’ (Cairncross, 2001), a large and growing collection of empirical evidence has revealed that electronic communication modes, such as telephone and emails, also tend to decline with longer spatial and relational distances, though not as sharply as that for face-to-face contacts (e.g., Carrasco et al., 2008; Kellerman, 2016; Loo, 2012; Mok et al., 2010; Tillema et al., 2010). A recent survey in Toronto has found that people tend to make more telephone calls to others within 100 miles,

H1. Spatial and relational distances still matter in the social networks of urban residents in OSNs. However, the roles of such distances vary among individuals of different socioeconomic (such as gender, age and education) and Internet use experience characteristics. Moreover, recent research on the e-working and e-shopping behaviours of urban residents indicates that these e-activity behaviours may be historically and spatially contingent (Dijst, Farag, & Schwanen, 2008; Loo & Wang, 2017b; Ren & Kwan, 2009). However, very little has 2

Cities xxx (xxxx) xxx–xxx

B. Wang et al.

proper has a higher level of land-use mix and accessibility to a wide variety of urban facilities in most Chinese cities such as Nanjing (Yuan, Gao, & Wu, 2016). As previously stated, researchers have found a positive correlation between local interactions, social capital and the built environment with mixed-use amenities (Cabrera & Najarian, 2015; Loo, Lam, Mahendran, & Katagiri, 2017). Similarly, we expect to find the existence of different spatial roles in OSNs among urban residents living in the city proper and residents living in the suburban areas.

been written in the urban scholarship to document the impact of historical and spatial settings, such as culture, experience and living environment, on the interpersonal interactions of urban residents in OSNs, especially in China. Social networks have always been highlighted in China studies (Chen & Sun, 2006; Davies, Leung, Luk, & Wong, 1995; Xin & Pearce, 1996). Under state socialism, social networks in both rural and urban living are mainly produced and developed at the local scale in traditional China (e.g., Chai, Xiao, Liu, & Ta, 2016; Fei, 1983; Hazelzet & Wissink, 2012; Ruan, Freeman, Dai, Pan, & Zhang, 1997; Zhang et al., 2017). Owing to geographical isolation, the lack of public transportation and the poor accessibility of public services in rural areas, rural residents are thought to have a greater sense of social responsibility than urban residents (Hofferth & Iceland, 1998). Generally, the acquaintance society in Chinese rural areas was characterised by strong ties that were often built around the village within which rural peasants help one another for goods and services (Fei, 1983). Economic reforms introduced since 1978 have triggered the rapid urbanisation process in which numerous rural residents left their farms to work in the city. These mobile workers usually do not hold a local hukou, and due to fewer interactions with locals and acquaintances in the city in their real life (Liu, Li, & Breitung, 2012; Xu & Chan, 2011) and their strong social ties with townsfolk in their hometowns, the virtual social networks of rural-urban migrants may differ from that of non-migrants. Therefore, the following hypothesis is set:

H4. Spatial distance plays a bigger role in the virtual interpersonal interactions of urban residents living in the city proper than in that of urban residents in the suburban areas.

3. Research design This research selected Nanjing, an ancient capital city of China in the core area of the Yangtze River Delta, as a case. It is the capital city of Jiangsu Province and covers an area of 4732 km2 with a population of approximately 8 million in 2010 according to the Sixth National Census. Nanjing is representative of the rapidly growing and large coastal cities of China; more importantly, it was once a socialist industrial city that accommodated a number of large-scale owned enterprises (Yuan et al., 2016). These enterprises left many Danwei communities, and some still exist in contemporary Nanjing. Moreover, Nanjing has a large number of rural migrants in recent decades (Yuan et al., 2016), resulting in rapid urban land expansion. Thus, Nanjing is deemed a suitable case study to study the dynamic and rapidly changing social networks in the broader context of urban transformation. With approximately 270 million active users in 2012 (CNNIC, 2012), Sina Weibo was selected as the OSNs for our research study. It was also the first and most popular OSNs in China.3 The data of this study were mined from the online information of Sina microbloggers. Given the need to conduct interviews about their socioeconomic and Internet use experience characteristics and other attributes related to urban transformation (i.e., Nanjing hukou/rural migrants with other cities' hukou, Danwei community/non-Danwei community and living in the city proper/suburban), we selected a total of 300 Sina microblogggers in December 2012 with the following sampling procedures, adopted from Zhen et al. (2016). First, participants were filtered by the geographical location in their user profiles, with Nanjing City being the filter criterion used to search for users on the Sina Weibo site. The site returned a list of around 1.2 million users within the specified geographical location and randomly listed them in web pages. The first 100 web pages were publicly available, each of which contained 20 users, with no more than six users selected from each page. The selected users had to meet two conditions: (1) the user had to be a regular adult user but not a celebrity, as a celebrity's OSNs functions more similarly to an advertisement platform than as a platform for making friends; and (2) the user had to be relatively active on OSNs with ≥400 followers and 400 followings and no less than six tweets per day. Each selected user was contacted for consent to participate in an online interview to gather relevant data. Failure to meet these requirements or refusal to participate in the interview resulted in exclusion from the study. Subsequent users were considered one by one until one was found to fulfil the above conditions. When no user in the same web page met the requirements, the next web page was explored until all 300 valid samples were recruited. The relationships between Sina microbloggers fall into three types, namely, follower, following and friend. Follower and following denote a unidirectional relationship between Sina microbloggers. For example, even if A follows B, B does not necessarily follow A. In this case, A is one

H2. Spatial and relational distances play weaker roles in the virtual interpersonal interactions of rural-urban migrants than in those of nonmigrants. In urban areas, social networks have mainly been developed within the ‘walled’ and ‘gated’ boundaries of work units, or Danwei, within which workplace, residence and social facilities were integrated under Maoist rule (Chai et al., 2016; Hazelzet & Wissink, 2012; Whyte & Parish, 1985). Various studies have found that the neighbourhoodbased life patterns in Danwei gave rise to mutual understanding and friendships indicative of close social ties among colleagues (Chai et al., 2016; Whyte & Parish, 1985). However, the transition from communism to market-socialism has destroyed the central role of Danwei as a site of social organisation for most urbanites (Chai et al., 2016; Whyte & Parish, 1985). The increased residential mobility and changing lifestyle during the urban transformation caused the intensive physical interactions with locals and acquaintances of the 1970s to subside (Chai et al., 2016). However, noting the differences between the social networks of different community types (Forrest & Yip, 2007; Hazelzet & Wissink, 2012) is important as well. Several studies suggest that urban residents living in a Danwei community tend to hold stronger neighbourhood relationships in their daily lives than do residents in nonDanwei communities (Forrest & Yip, 2007), indicating that the strong social ties among locals and acquaintances have not disappeared. It is reasonable to consider how the physical social networking experiences of residents impact their online interpersonal interactions. Thus, we set up the following hypothesis: H3. Spatial and relational distances have stronger roles in the virtual interpersonal interactions of urban residents living in a Danwei community than in those of urban residents in a non-Danwei community. Additionally, China's urbanisation process has long been criticised for its rapid land expansion and residential segregation in most cities (Li & Wu, 2008; Wu, 2015). In suburban areas, gated communities prevail and have dominated commercial housing projects (Wu, 2010). The rise of gated communities, compounded by the lack of public services, has led to fewer opportunities for physical interpersonal interactions among residents who live nearby (Hazelzet & Wissink, 2012). Close social networks among people living nearby and locals have also been negatively impacted by the rapid urban expansion and rehousing projects in China (Wu & He, 2005). However, unlike suburban areas, the city

3 The survey report on the first year of Weibo use in China (Zhongguo weibo yuannian shichang baipishu). Accessed 29 December 2014, from http://www.slideshare.net/ lxm19871231/ss-5245420.

3

Cities xxx (xxxx) xxx–xxx

B. Wang et al.

of B's followers and can read, forward and reply to each of B's tweets; B is then classified as one of A's followings. If A and B follow each other, then they are both classified as each other's friends. A friend relationship tends to be stronger than the follower and following relationships (Zhen et al., 2016). In the interview, respondents were first asked about their socioeconomic characteristics, including gender, age, education, employment status and Internet use experience characteristics, such as smartphone ownership and the daily time spent on the Internet. To measure the impacts of urban transformation, respondents were asked to answer (1) whether they hold a Nanjing hukou, (2) whether they live in a Danwei community and (3) whether they live in the city proper. Respondents were then asked to provide the percentage of local users who also live in Nanjing4 and the acquaintances in their followers, followings and friends whom they already know in real life. Note that a relationship between users does not necessarily lead to actual interpersonal interactions between them in Sina Weibo. Hence, respondents were asked to check and count the actual times they interacted with other users in Sina Weibo in the past two weeks. The interactions included (1) reply and forward tweets and (2) exchange messages. On the basis of these details, one can calculate the percentage of their interactions with local users and acquaintance users. These two percentages were used in natural logarithms (i.e., LNInter_NJ and LNInter_Acq) as the dependent variables in our analysis later to examine the roles of spatial distance and relational distance in their online interpersonal interactions. It is understood that our samples cannot represent the entire urban population of Nanjing, and the findings and discussions are based on the samples collected in this study. Table 1 provides some descriptive statistics of the samples. Overall, 65.00% of them were aged below 30, which aligns with the popular use of OSNs among the youth of China (CNNIC, 2012) and other countries (Thulin & Vilhelmson, 2005). A majority of them owned smartphones (84.33%). This finding can be explained by the high penetration of Weibo among people with mobile phones (CNNIC, 2012), which also explains the high percentage of highly educated people observed in our samples. An expectedly low rate of respondents living in Danwei community represents the transition from communism to market-socialism in contemporary China. Most of the samples were employed and were holding Nanjing hukou. In addition, about half of them spent more than 1.5 h using the Internet.

Table 1 Sample profile. Variables

Categories

Case

Percentage

Gender

Male Female 18–19 21–29 30–39 40 or above High school or below Junior college Bachelor Master or above Employed Non-employed Rural migrants with other cities' hukou Nanjing hukou Danwei community Non-Danwei Community Living in the city proper Living in suburban areas Smartphone ownership No smartphone ownership < 1.5 h 1.5–3.0 h > 3.0 h

155 145 51 144 73 32 25 84 137 54 253 47 225

51.67 48.33 17.00 48.00 24.33 10.67 8.33 28.00 45.67 18.00 84.33 15.67 75.00

75 61 239 201 99 253 47 123 92 85

25.00 20.33 79.67 67.00 33.00 84.33 15.67 41.00 30.67 28.33

Age

Education

Employment status Nanjing hukou

Community type City proper Smartphone ownership Daily Internet use time

et al., 2005; Takhteyev et al., 2012). Interestingly, the combined percentage of friends in Nanjing and the other cities of Jiangsu Province (49.6%) is higher than the percentage of followers (49.6%) and followings (45.6%). The opposite is true for the other provinces (including bordering and non-bordering provinces) with the percentage of followers (15.3% for bordering provinces and 35.1% for non-bordering provinces) and followings (14.7% for bordering provinces and 34.7% for non-bordering provinces) higher than friends (13.5% for bordering provinces and 26.4% for non-bordering provinces). This striking pattern reveals that urban residents tend to build more virtual relationships with locals as well because, as explained earlier, a friend relationship is stronger than the follower and following relationships in Sina Weibo (Zhen et al., 2016). Moreover, we found that 43.1% of followers, 44.4% of followings and 52.1% of friends are acquainted in real life. This high percentage of relationships with acquaintances in our samples among Sina Weibo users is consistent with other studies in Western countries (Thulin & Vilhelmson, 2005). It further confirms that the Internet is mainly used for communication with people already known from real life (Thulin & Vilhelmson, 2005; Tillema et al., 2010). Again, a higher percentage of acquaintances in real life is found among friends than with followers and following, an outcome that can be attributed to a closer relational distance in friend relationships in OSNs. Overall, spatial and relational distances still matter in explaining virtual relationships, especially friendships in OSNs. In addition, the high percentage of relationships in Sina Weibo with local and acquaintance users suggests that OSNs may play an important role in maintaining and strengthening existing relationships with others in real life in China.

4. Results 4.1. Building relationships with local and acquaintance users in Sina Weibo Table 2 shows a dominance of relationships in Sina Weibo among users within the same locality. Specifically, 35.5% of followers, 37.8% of followings and 49.7% of friends of the 300 samples were living in Nanjing. Within Jiangsu Province, the difference between Nanjing and the other cities in Jiangsu Province is striking, with the average corresponding percentages for the other cities at 1.2%, 0.7% and 0.9%, only respectively. The percentage difference between Jiangsu Province and the other 30 provinces is also ostensible. While 49.6% of the followers, 45.6% of the followings and 60.1% of the friends documented came from Jiangsu Province, the corresponding average percentages for the other 30 provinces in China were 1.2%, 1.3% and 0.9%, respectively. Among these other 30 provinces, about twice as much of followers, followings and friends lived in bordering provinces versus nonbordering provinces. The above findings clearly suggest a spatial distance decay pattern in relationship building in OSNs. In other words, relationships in OSNs also tend to be affected by spatial proximity with stronger online relationships among people living nearby (Liben-Nowell

4.2. Interacting with local and acquaintance users in Sina Weibo Owing to the difficult and time-consuming nature of differentiating the situation among followers, followings and friends, respondents were only asked to report the overall percentage of their interactions in Sina Weibo with local users and acquaintance users. In other words, respondents were not required to differentiate between their interactions among followers, followings and friends. Specifically, the corresponding mean values of the overall percentage were 51.5% and 56.9%, respectively. Both percentages were higher than the corresponding percentage of relationships in Sina Weibo. Based on the descriptive

4 This information can be obtained from the geographical location in the users' profiles. A simple website crawler application was used to automatically collect these details to guarantee the accuracy of the data.

4

Cities xxx (xxxx) xxx–xxx

B. Wang et al.

Table 2 Percentage of followers, followings and friends in Nanjing, other parts of Jiagsu Province, Bordering Provinces and Non-bordering Provinces in China for 300 adult Sina microbloggers in Nanjing. Nanjing (%)

Other parts of Jiangsu Province (%)

Bordering provinces (%)

Non-bordering provinces in China and others (%)

I

II

III

I

II

III

I

II

III

I

II

III

35.5

37.8

49.7

14.1 (1.2)

7.8 (0.7)

10.4 (0.9)

15.3 (3.8)

14.7 (3.7)

13.5 (3.4)

35.1 (1.2)

39.7 (1.3)

26.4 (0.9)

Note: I = Followers; II = Followings; III = Friends. Figures in parentheses represent the corresponding average percentage for each city/province.

analysis, urban residents were not only more likely to build relationships with local and acquaintance users but also to interact with them more frequently in Sina Weibo. To further test our hypotheses, two robust regression models were developed to explore the variations in interpersonal interactions of Sina microbloggers with local and acquaintance users, respectively. The dependent variable is the percentage of interpersonal interactions with the respective users in real life. Note that a Sina microblogger's higher levels of interaction with local and acquaintance users may be attributed to their higher percentage of relationships with these users. Therefore, the percentage of friend relationships Sina microbloggers had with local and acquaintance users in natural logarithms (i.e., LNFri_NJ and LNFri_Acq) were introduced as the control variables in the regression model. Prior to estimating the regression models, a Pearson correlation analysis was used to check the correlation between independent variables. All correlation coefficients among explanatory variables in the final model were < 0.33 and/or statistically insignificant. A variance inflation factor test was also conducted, and no evidence of multicollinearity among independent variables was found. Moreover, the results of the Breusch–Pagan test and the White test revealed the existence of heteroscedasticity. Thus, the robust regression model was adopted to eliminate the effect of heteroscedasticity. Generally, this approach yields asymptotically valid results (White, 1980). Table 3 summarizes the results of the robust regression model exploring the interactions of urban residents with local users (Model 1) and acquaintance users (Model 2) in Sina Weibo. The significant effects of several independent variables prove that the interactions between urban residents and local and acquaintance users in OSNs are significantly influenced by their socioeconomic and Internet use experience characteristics. In line with the descriptive analysis previously discussed, the results support the first hypothesis. To reiterate, spatial and relational distances still matter in the social networks of urban residents in OSNs. However, the roles of such distances vary among individuals of different socioeconomic (such as gender, age and education) and Internet use experience characteristics. As expected, although Sina Microbloggers who have a higher percentage of relationship with local (LNFri_NJ) and acquaintance users (LNFri_Acq) tend to have more personal interactions with these users, this tendency is statistically insignificant. This finding corroborates the observations that a relationship between users in online social networks does not directly lead to more actual interpersonal interactions (Laniado et al., 2017). Generally, male respondents were significantly more likely to have interpersonal interactions with local and acquaintance users. Based on earlier findings, active OSNs users usually have more relationships and interactions with others online (Huang & Sun, 2014). Furthermore, as female urban residents have a greater need to save time and reduce travel because they experience more stringent space–time constraints (Loo & Wang, 2017b; Ren & Kwan, 2009), they tend to be more engaged in online activities and thus have more opportunities to interact with other users apart from those in the same city or existing acquaintances. This finding is in line with the strongly negative effects of Internet use time. The more time one spends on the Internet, the more likely he/she will have frequent interactions with users who are non-locals and strangers. Note that those who have spent lots of time using the Internet

Table 3 Robust regression results for the role of spatial and relational distances in the personal interactions of urban residents in Sina Weibo. Variable

Control variables LNFri_NJ LNFri_Acq Socioeconomic variables Gender (female = ref.) Age (18–19 = ref.) 20–29 30–39 40 or above Education (master or above = ref.) High school or below Junior college Bachelor Employment (nonemployed = ref.) Internet use experience variables Smartphone ownership (no smartphone = ref.) Daily Internet use time (< 1.5 h = ref.) 1.5–3.0 h > 3.0 h Urban transformation variables Nanjing hukou (rural migrants with other cities' hukou = ref.) Community type (nonDanwei community = ref.) City proper (living in the suburban = ref.) Adjusted R2 F-value

Model 1

Model 2

Coef.

Robust S.E.

0.206

0.115

Coef.

Robust S.E.

0.036 0.103

⁎⁎⁎

⁎⁎

0.019

0.125

0.097 ⁎⁎⁎



0.018

0.192 0.185⁎⁎ 0.001

0.074 0.061 0.039

0.165 0.232⁎⁎⁎ 0.226⁎⁎⁎

0.071 0.058 0.047

0.075⁎ 0.059⁎ 0.049⁎ −0.036

0.036 0.029 0.024 0.032

0.027 0.025 0.039 0.007

0.039 0.029 0.026 0.031

−0.054

0.027

−0.012

0.027

−0.186⁎⁎⁎ −0.270⁎⁎⁎

0.023 0.031

−0.149⁎⁎⁎ −0.269⁎⁎⁎

0.022 0.031

0.259⁎⁎

0.023

0.108⁎⁎⁎

0.022

0.110⁎⁎⁎

0.023

0.134⁎⁎⁎

0.023

0.052⁎⁎

0.019

0.006

0.019

0.610 32.13

0.617 33.10



p < 0.05. p < 0.01. ⁎⁎⁎ p < 0.001. ⁎⁎

are more likely to join specific interest communities where they actively talk to others whom they have never met before. As one respondent who spends > 3.0 h a day on the Internet added, ‘To some extent, Sina Weibo has become an important platform for me to get news, especially entertainment and fashion news. And I like to discuss these with my cyberfriends even though we did not know one another before Sina Weibo. What matters is that we have the same interests and have fun in our conversations. It does not matter where they come from and whether we will meet in the real world in the future’. As such, OSNs play an important role in extending their social relationships online. For Sina microbloggers who used the OSNs largely for sharing and exchanging information regarding specialised interests, the roles of spatial and relational distances become relatively weaker. However, no evidence that such extensions of their online social networks actually generate more physical or real-life interactions has been found in our dataset. 5

Cities xxx (xxxx) xxx–xxx

B. Wang et al.

a short period. And after work, we have less interaction; it is hard to say that we could become friends in real life’. However, OSNs can present important channels of communication in the migrants' social network. In view of their greater sense of responsibility to others in their hometown, rural-urban migrants actively share related employment information with users from their hometown online. ‘I got the job information from one of my fellow villagers who moved to Nanjing early. Generally, it is not easy for migrants to get job information in large cities, and they become unemployed as a result, so I would not hesitate to share information on new jobs with my villagers. Sharing the information in Sina Weibo is free and I can enable many of my fellow villagers, especially those who still stayed in my hometown, to see the information simultaneously in a timely manner’. This finding further suggests that OSNs may also influence migration aspirations and decision-making (Dekker et al., 2016). Moreover, it seems that residents without hukou status usually tend to show very little interest in local news. As one migrant adds, ‘Hometowns give you a feeling of familiarity, and I like to make friends and interact with users in my hometown through Sina Weibo. More importantly, I can get timely news of my hometown from them, and sometimes these are hard to get from local newspapers or TV here in Nanjing’. As such, OSNs may also function to help create social capital for migrants in big cities in China (Liu et al., 2012). Respondents living in a Danwei community are more likely to interact with local and acquaintance users. Once again, this may be explained by their common experience of building social networks with neighbours who live and work together in ‘walled’ and ‘gated’ compounds in their daily lives (Chai et al., 2016; Hazelzet & Wissink, 2012; Whyte & Parish, 1985). As one respondent living in a Danwei community added, ‘First, through Sina Weibo, I have built friendships with almost everyone living in the same Danwei community who also uses the platform, and we interact with one another quite often. Second, I really like engaging in interactions with other local and acquaintance users, and these interactions seem to be more real for me’. As such, OSNs can play a key role in strengthening neighbourhood relationships as well. Another young respondent commented that ‘due to the reform of the Danwei community, the Danwei no longer organises as many community activities as it did when I was a child. However, we now sometimes share or discuss our community issues and collect opinions on organising some community activities through Sina Weibo’. In this way, people living in a Danwei community maintain strong ties with former neighbours through the OSNs. Third, respondents living in the city proper are more likely to interact with local users in Sina Weibo. One reason may be related to the high concentration of urban public and commercial facilities in the city proper of Nanjing (Yuan et al., 2016). Using these facilities increases the chance of building multiple social networks with locals in the real world, and it is possible for social networks in the real world to extend to the virtual social networks in OSNs (Mok et al., 2010; Thulin & Vilhelmson, 2005). Another factor may be the low frequency of physical interactions among neighbours in gated communities in suburban areas. As one respondent living in one gated community commented, ‘I have been in this community for about three years; however, I do not even know the name of my neighbour. For me, I usually choose to stay at home after work because there are almost no places worth visiting in our neighbourhood. Browsing Sina Weibo is a good option to kill time. Even though I have some local friends in Sina Weibo, very few of them live close to me. After I moved here, I gradually have fewer interactions with my local friends’. Our finding further suggests that the built environment may also influence virtual interpersonal interactions. Therefore, it is important to consider spatial planning when aiming to promote local interactions in OSNs.

Moreover, no obvious difference was observed between the percentage of interpersonal interactions of people with or without a smartphone and the local/non-local users and acquaintance/stranger users in Sina Weibo. As some scholars argued, having a smartphone is more often a consideration for people's face, or mianzi, than the technology contained in the phone (Chu, 2008). Nowadays, instead of studying hether or not people use smart ICTs devices, the extent to which they use these devices may be more relevant to a better understanding of the effects of technology on the e-activity behaviours of urban residents (Loo & Wang, 2017b), such as their interactions in OSNs. Compared with respondents aged 18–19 years, respondents in their 20s, 30s and 40s or above are more likely to interact with users they already know in real life. Respondents in their 20s and 30s likewise tend to have more interactions with local users; however, the tendency for respondents in their 40s or above is statistically insignificant. This outcome may be related to the perception that talking to strangers online in OSNs is a ‘childish’ activity popular among the youth (Thulin & Vilhelmson, 2005). Similarly, these age-related effects may imply that OSNs have been used mostly by teenagers to extend their virtual social networks, rather than the older age groups. Teenagers often have relatively smaller social networks in the real world and hold a more participative and open attitude towards OSNs (Dekker et al., 2016; Huang & Sun, 2014). As one typical teenage respondent highlighted, ‘I like to follow other users and I also like to be followed by other users in Sina Weibo. It is cool if you have more followers or friends, and it makes you feel really popular and important. And it is fine for me to contact anyone in Sina Weibo without thinking too much about who they are and whether I already know them’. On the contrary, Sina Weibo seems relatively new for respondents in their 40s or above. Some of them chose to use Sina Weibo because of recommendations from their friends; and thus, they interact with other users more carefully online. The corresponding significantly positive effect of age in Model 2 demonstrates that users in their 40s or above tend to have more interactions with users they already know in real life. However, as those people usually have a larger social network that spans beyond locals in the real world (Tillema et al., 2010), the role of spatial distance is not statistically obvious. Also, employed respondents were not found to have significantly more interactions with local and acquaintance users. However, compared with urban residents with a master's degree or above, those with a lower education level have a higher possibility of having more interactions with local users. Again, the corresponding insignificant effect of education level in Model 2 suggests that people with a higher education level may have a social network that spans across a bigger space in their real lives (Mack, Marie-Pierre, & Redican, 2017). This finding also aligns with earlier observations that education helps enrich one's social networks, thereby increasing the possibility of success in one's future development (Knack & Keefer, 1997). However, it is also important to highlight the possible capability of OSNs to extend one's social network in their daily lives (Dekker et al., 2016). It is interesting to highlight how significantly variables that relate to urban transformation affect the interactions of urban residents with local and acquaintance users in Sina Weibo. The findings support Hypotheses 2–4, which indicate that the interpersonal interactions of urban residents in OSNs are also been heavily influenced by the broader context of urban transformation in contemporary China. First, compared to rural-urban migrants with other cities' hukou, urban residents with Nanjing hukou status are significantly more likely to have frequent interactions with local and acquaintance users. This finding may be attributed to their weak ties with local people in their real lives, given that it takes time for migrants to build up social capital in their new location (Hofferth & Iceland, 1998). As one migrant respondent said, ‘Although I have worked in Nanjing for many years, I cannot find my roots here. As you may know, it is common to meet and work with others during my life in Nanjing; however, we usually stay together for

5. Conclusion and discussion This work analyses the roles played by spatial and relational 6

Cities xxx (xxxx) xxx–xxx

B. Wang et al.

variables related to urban transformation that may influence electronic communication behaviours. As a pioneer study, we did not consider other potential variables such as attitudinal characteristics, which have been proven to significantly influence e-activity behaviours (Loo & Wang, 2017b).

distances in an OSNs platform, Sina Weibo, in China. The Chinese culture heavily emphasises social networks. Our research shows that urban residents tend to build virtual relationships more strongly with local and acquaintance users than those separated by longer spatial distance and less close relational distance. Spatial and relational distances still clearly affect social networks in OSNs. However, the roles of such distances vary among individuals of different socioeconomic and Internet use experience characteristics. Furthermore, this study points out the importance of different urban contexts on the roles of spatial and relational distances by examining the variations in interpersonal interactions between Sina microbloggers and their followers, followings and friends locally (model 1) and those who they already know in reallife (model 2). The results shed important lights on the debate about the relationship between and the mutual influence of social networks in the virtual space and in the real world (Huang & Sun, 2014; Mok et al., 2010; Sgroi, 2008). Overall, there is clear evidence that physical and relational distances still play leading roles in shaping virtual relationships and interactions in the e-society (Loo, 2012). Methodologically, a valuable contribution of this study is to show that the combination of online data records and interviews is important and applicable for future urban studies. On the one hand, data in the existing studies were gathered mainly through questionnaires, interviews and communication diaries, which depended heavily on the cooperation of respondents and the respondents' memories (e.g., Mok et al., 2010; Mok & Wellman, 2007; Thulin & Vilhelmson, 2005; Tillema et al., 2010). On the other hand, some recent studies have begun to use very big datasets left by OSNs users online (e.g., Liben-Nowell et al., 2005; Takhteyev et al., 2012), which have also been challenged for their deficient socioeconomic characteristics in data collection from the online profiles of users (Graham & Shelton, 2013). However, our study shows that it is possible and valuable to combine the ‘big data’ of online traces and ‘small data’ of interviews in one study. As shown in this study, the online records can provide an accurate historical account of the number of relationships and frequencies of interactions over a period of time; while the interviews provide valuable information about personal attributes and views. Through combining the online data records and interviews, this study represents an innovative, though limited, attempt to analyse the effects of informatisation and urban transformation on the social networks of urban residents. The results also have implications for policy making. First, OSNs can be a useful tool for local governments to support migrants. For example, a specific Weibo group aimed at attracting migrants from the same province/city/town provides opportunities for them to interact and, more importantly, help one another online. This group helps enhance communications in the migration networks and thus maintain strong ties among the migrants. Second, it is important to consider the influence of the built environment on OSNs. As our study indicates, OSNs sometimes function to maintain and extend one's social networks in the real world, but a built environment with a higher level of land-use mix and accessibility to a wide variety of urban facilities can also contribute to more local interpersonal interactions among people who are also connected by OSNs. Recently, many ‘community construction’ policies have been proposed in China to counter the social problems caused by the rapid urban land expansion and residential segregation in cities. Although OSNs can provide additional channels for communication among neighbours, it is still important to take spatial planning seriously because virtual interactions are not independent of the local built environment. Some avenues for further research are also identified. Our online interviews only required respondents to distinguish members with whom they had interacted with in the past two weeks, either from within Nanjing or from other cities (model 1), as well as those they knew or did not know in real life (model 2). If possible, a more detailed classification of spatial distance and a more precise definition of relational distance could describe the role of spatial and relational distances in OSNs more precisely. Further studies may also probe into other

Acknowledgements This paper was funded by the National Natural Science Foundation of China (41571146, 41671154). References Amin, A., & Thrift, N. (2002). Cities: Reimagining the urban. UK: Polity Press. Cabrera, J. F., & Najarian, J. C. (2015). How the built environment shapes spatial bridging ties and social capital. Environment and Behavior, 47(3), 239–267. Cairncross, F. (2001). The death of distance: How the communications revolution is changing our lives. Boston: Harvard Business Press. Carrasco, J., Miller, E., & Wellman, B. (2008). How far and with whom do people socialize?: Empirical evidence about distance between social network members. Transportation Research Record: Journal of the Transportation Research Board, 2076, 114–122. Chai, Y. W., Xiao, Z. P., Liu, T. B., & Ta, N. (2016). Danwei perspective on urban China. Nanjing: Southeast University Press (In Chinese). Chen, X., & Sun, J. (2006). Sociological perspectives on urban China: From familiar territories to complex terrains. China Information, 20(3), 519–551. Chu, W. C. R. (2008). The dynamics of cyber China: The characteristics of Chinese ICT use. Knowledge, Technology, and Policy, 21(1), 29–35. CNNIC (2012). The 29th survey on China Internet development in China. Beijing: China Internet Network Information Center (In Chinese). Davies, H., Leung, T. K., Luk, S. T., & Wong, Y. H. (1995). The benefits of “Guanxi”: The value of relationships in developing the Chinese market. Industrial Marketing Management, 24(3), 207–214. Dekker, R., Engbersen, G., & Faber, M. (2016). The use of online media in migration networks. Population, Space and Place, 22(6), 539–551. Dijst, M., Farag, S., & Schwanen, T. (2008). A comparative study of attitude theory and other theoretical models for understanding travel behaviour. Environment and Planning A, 40(4), 831–847. Fei, X. (1983). Rural development in China: Prospect and retrospect. Chicago: University of Chicago Press. Fischer, C. S. (1982). To dwell among friends: Personal networks in town and city. Chicago: University of Chicago Press. Forrest, R., & Yip, N. M. (2007). Neighbourhood and neighbouring in contemporary Guangzhou. Journal of Contemporary China, 16(50), 47–64. Glaeser, E. (2011). Triumph of the city: How our greatest invention makes us richer, smarter, greener, healthier, and happier. London: Penguin. Graham, M., & Shelton, T. (2013). Geography and the future of big data, big data and the future of geography. Dialogues in Human Geography, 3(3), 255–261. Hazelzet, A., & Wissink, B. (2012). Neighborhoods, social networks, and trust in postreform China: The case of Guangzhou. Urban Geography, 33(2), 204–220. Hofferth, S. L., & Iceland, J. (1998). Social capital in rural and urban communities. Rural Sociology, 63(4), 574–598. Huang, R., & Sun, X. (2014). Weibo network, information diffusion and implications for collective action in China. Information, Communication & Society, 17(1), 86–104. Kaplan, A. M., & Haenlein, M. (2010). Users of the world, unite! The challenges and opportunities of Social Media. Business Horizons, 53(1), 59–68. Kearns, A., & Forrest, R. (2000). Social cohesion and multilevel urban governance. Urban Studies, 37(5–6), 995–1017. Kellerman, A. (2016). Geographic interpretations of the internet. Berlin: Springer. Knack, S., & Keefer, P. (1997). Does social capital have an economic payoff? A crosscountry investigation. The Quarterly Journal of Economics, 112(4), 1251–1288. Laniado, D., Volkovich, Y., Scellato, S., Mascolo, C., & Kaltenbrunner, A. (2017). The impact of geographic distance on online social interactions. Information Systems Frontiers. http://dx.doi.org/10.1007/s10796-017-9784-9. Li, Z. G., & Wu, F. L. (2008). Tenure-based residential segreggation in post-reform Chinese cities: A case study of Shanghai. Transactions of the Institute of British Geographers, 33(3), 404–419. Liben-Nowell, D., Novak, J., Kumar, R., Raghavan, P., & Tomkins, A. (2005). Geographic routing in social networks. Proceedings of the National Academy of Sciences of the United States of America, 102(33), 11623–11628. Liu, Y., Li, Z. G., & Breitung, W. (2012). The social networks of new-generation migrants in China's urbanized villages: A case study of Guangzhou. Habitat International, 36(1), 192–200. Loo, B. P. Y. (2012). The E-society. New York: Nova Science. Loo, B. P. Y., Lam, W. W., Mahendran, R., & Katagiri, K. (2017). How is the neighborhood environment related to the health of seniors living in Hong Kong, Singapore, and Tokyo? Some insights for promoting aging in place. Annals of the American Association of Geographers, 107(4), 812–828. Loo, B. P. Y., & Wang, B. (2017a). Progress of e-development in China since 1998. Telecommunications Policy, 41(9), 731–742. Loo, B. P. Y., & Wang, B. (2017b). Factors associated with home-based e-working and eshopping in Nanjing, China. Transportation. http://dx.doi.org/10.1007/s11116-017-

7

Cities xxx (xxxx) xxx–xxx

B. Wang et al.

Implications for the social integration of migrants. Urban Studies, 53(4), 651–668. Wellman, B., & Leighton, B. (1979). Networks, neighborhoods, and communities approaches to the study of the community question. Urban Affairs Review, 14(3), 363–390. White, H. (1980). A heteroskedasticity-consistent covariance matrix estimator and a direct test for heteroskedasticity. Econometrica: Journal of the Econometric Society, 817–838. Whyte, M. K., & Parish, W. L. (1985). Urban life in contemporary China. Chicago: University of Chicago Press. Wu, F. (2010). Gated and packaged suburbia: Packaging and branding Chinese suburban residential development. Cities, 27(5), 385–396. Wu, F. (2015). Planning for growth: Urban and regional planning in China. London: Routledge. Wu, F., & He, S. J. (2005). Changes in traditional urban areas and impacts of urban redevelopment in three neighbourhoods in Nanjing, China. Tijdschrift voor Economische en Sociale Geografie, 96(1), 75–95. Xin, K. K., & Pearce, J. L. (1996). Guanxi: Connections as substitutes for formal institutional support. Academy of Management Journal, 39(6), 1641–1658. Xu, Y., & Chan, E. H. W. (2011). Community question in transitional China, a case study of state-led urbanization in Shanghai. Journal of Urban Planning and Development, 137(4), 416–424. Yuan, F., Gao, J., & Wu, J. (2016). Nanjing-an ancient city rising in transitional China. Cities, 50, 82–92. Zhang, M., Wu, W., Zhong, W., Zeng, G., & Wang, S. (2017). The reshaping of social relations: Resettled rural residents in Zhenjiang, China. Cities, 60, 495–503. Zhen, F., Wang, B., & Chen, Y. X. (2016). Research on China's city network based on users' friend relationships in online social networks: A case study of Sina Weibo. GeoJournal, 81(6), 937–946.

9792-0. Mack, E. A., Marie-Pierre, L., & Redican, K. (2017). Entrepreneurs' use of internet and social media applications. Telecommunications Policy, 41(2), 120–139. Mok, D., & Wellman, B. (2007). Did distance matter before the Internet?: Interpersonal contact and support in the 1970s. Social Networks, 29(3), 430–461. Mok, D., Wellman, B., & Carrasco, J. (2010). Does distance matter in the age of the Internet? Urban Studies, 47(13), 2747–2783. Ren, F., & Kwan, M. P. (2009). The impact of geographic context on e-shopping behavior. Environment and Planning. B, Planning & Design, 36(2), 262–278. Ruan, D., Freeman, L. C., Dai, X., Pan, Y., & Zhang, W. (1997). On the changing structure of social networks in urban China. Social Networks, 19(1), 75–89. Sgroi, D. (2008). Social network theory, broadband and the future of the World Wide Web. Telecommunications Policy, 32(1), 62–84. Sharmeen, F., Arentze, T., & Timmermans, H. (2014). Dynamics of face-to-face interaction frequency: Role of accessibility, urbanization, changes in geographical distance and path dependence. Journal of Transport Geography, 34, 211–220. Takhteyev, Y., Gruzd, A., & Wellman, B. (2012). Geography of Twitter networks. Social Networks, 34(1), 73–81. Thulin, E., & Vilhelmson, B. (2005). Virtual mobility of urban youth: ICT-based communication in Sweden. Tijdschrift voor Economische en Sociale Geografie, 96(5), 477–487. Tillema, T., Dijst, M., & Schwanen, T. (2010). Face-to-face and electronic communications in maintaining social networks: The influence of geographical and relational distance and of information content. New Media & Society, 12(6), 965–983. Wang, B., Zhen, F., Xi, G. L., Qian, Q., Wu, C. Y., & Zhang, H. (2013). A study of cybergeography based on micro-blog users' relationship: With a case of Sina micro-blog. Geographical Research, 32(2), 380–391. Wang, Z., Zhang, F., & Wu, F. (2016). Intergroup neighbouring in urban China:

8