Can Social Capital Shorten the Social Distance of Urban Residents? Evidence From the Spatial Stratification in China

Abstract

As the commodification of urban housing increases, the phenomenon of class segregation resulting from different types of housing becomes increasingly apparent. Research on the social distance between condominium owners and relocated residents contributes to understanding the social stratification logic in contemporary Chinese urban areas. Using sampled data from the 2019 China Urban Resident Living Space Survey, this study employs multiple linear regression and random forest models to explore the impact of online and offline social capital on the social distance between the two types of property ownership groups (i.e., condominium owners and relocated residents). The results show that both online and offline social capital have the potential to create new bridging social capital, facilitating heterogeneous social interactions. Additionally, online social capital effectively supplements offline social capital and significantly promotes the reduction of social distance between the two property ownership groups. Based on the study's findings, community governance should enhance the frequency and quality of interaction between condominium owners and relocated residents through offline community activities, resident organizations, and diverse online community platforms to promote social integration.

Full text

Sociology Compass
-
ORIGINAL ARTICLE
OPEN ACCESS
Can Social Capital Shorten the Social Distance of Urban
Residents? Evidence From the Spatial Stratication in
China
Ruolin Wang
1
| Yifei Li
1
| Kai Zhou
2
1
School of Population and Health, Renmin University of China, Beijing, China |
2
Policy Research Center for Environmental and Economic, Ministry of
Ecology and Environment of the People's Republic of China, Beijing, China
Correspondence: Yifei Li (wrl0415@ruc.edu.cn)
Received: 24 April 2024 | Revised: 11 April 2025 | Accepted: 5 May 2025
Funding: The authors received no specic funding for this work.
Keywords: community governance | random forest | social capital | social distance | social integration | spatial stratication
ABSTRACT
As the commodication of urban housing increases, the phenomenon of class segregation resulting from different types of
housing becomes increasingly apparent. Research on the social distance between condominium owners and relocated residents
contributes to understanding the social stratication logic in contemporary Chinese urban areas. Using sampled data from the
2019 China Urban Resident Living Space Survey, this study employs multiple linear regression and random forest models to
explore the impact of online and ofine social capital on the social distance between the two types of property ownership groups
(i.e., condominium owners and relocated residents). The results show that both online and ofine social capital have the po-
tential to create new bridging social capital, facilitating heterogeneous social interactions. Additionally, online social capital
effectively supplements ofine social capital and signicantly promotes the reduction of social distance between the two
property ownership groups. Based on the study's ndings, community governance should enhance the frequency and quality of
interaction between condominium owners and relocated residents through ofine community activities, resident organizations,
and diverse online community platforms to promote social integration.
1
|
Introduction
Social issues in transitional China have garnered considerable
attention. In the process of Chinese social transformation, cities
play a crucial role. With the continuous development of ur-
banization in China, the degree of land and housing commod-
ication has deepened, leading to spatial stratication as a new
trend in Chinese social stratication (Xu and Shi 2016). Ac-
cording to (Bourdieu (1984), 108), “there is an endless pursuit of
segmented elds in class society,” and the distribution and
segregation of urban space are widespread. In China, the
intrinsic characteristics of housing (e.g., property rights and
prices) and qualitative factors (e.g., area, quantity, functional
zoning, and geographical location) have created signicant
differences. This highlights the signicant role of housing in
indicating social status, giving rise to distinctly stratied hous-
ing status groups (H. Lu 2014; Li 2009).
Among the various divisions and segregations of groups, the
differences between commercial and relocation housing are
remarkable. Commercial housing refers to houses built by real
estate development companies for market sale, while relocation
housing refers to housing provided by developers to resettled
residents. To construct new commercial housing, existing
buildings on the site need to be demolished. Those who own the
original buildings in China are called resettled residents.
This is an open access article under the terms of the Creative Commons Attribution‐NonCommercial‐NoDerivs License, which permits use and distribution in any medium, provided the original work
is properly cited, the use is non‐commercial and no modications or adaptations are made.
© 2025 The Author(s). Sociology Compass published by John Wiley & Sons Ltd.
Sociology Compass, 2025; 19:e70071 1 of 13
https://doi.org/10.1111/soc4.70071

Generally, relocation housing is situated in the same commu-
nity as newly built commercial housing or may be separated by
only a wall. Regarding the differences between commercial and
relocation housing, rst, in terms of price, the selling price of
relocation housing is generally lower than that of commercial
housing. Second, with respect to property rights, relocation
housing cannot enjoy the same rights as commercial housing,
such as mortgage loans and listing, and is also subject to re-
strictions in transactions. Third, in terms of quality, factors such
as the orientation and layout of relocation housing are often
criticized.
In the process of rapid urbanization, the continuous prot‐
seeking behavior of capital has exacerbated this divergence,
and “the quality of urban life has become a commodity for the
rich” (Harvey 2014, 5). This stratication is further institution-
alized by China's hukou system, which historically ties social
welfare access to one's registered residency (urban or rural
hukou). Rural‐to‐urban migrants—often resettled households—
presumably retain rural hukou status even after relocation.
Hence, the property they own after urbanization may be limited
in circulation in the market. Consequently, hukou distinctions
compound housing commodication's segregated effects,
creating overlapping class and institutional boundaries within
mixed communities.
McPherson et al. (2001) indicate that homophily in social
interaction is a universal phenomenon that signicantly affects
subjective value judgments and social actions. Although
residing in the same community, commercial housing owners
often have higher levels of education, wages, and socioeconomic
status compared with relocated residents who receive policy
preferential treatment, leading to potential differences in their
value identication and a possibly signicant social distance
between the two groups, thereby triggering social conicts.
Social media breaks the homogeneity principle of traditional
social interaction. Netizens from different social classes
communicate through the internet, crossing the class barriers in
physical space, thereby reducing social distance between
different classes. The 27th China Internet Development Status
Statistical Report released by the China Internet Network In-
formation Center (CNNIC) shows that as of April 2020, the scale
of Chinese Internet users exceeded 900 million, making China a
country where residents are generally connected to the internet.
The virtual space constructed by the internet has become part of
people's lives, with local online social interactions affecting
people's ofine lives, generating a considerable amount of social
interaction known as “macro social capital” by scholars such as
Putnam (1993). For commercial housing owners and relocated
residents in the same community, the internet provides more
opportunities for interaction and mutual understanding,
increasing the community's macro social capital, which helps to
reduce social distance.
Therefore, this study seeks to answer the following question:
does online interaction contribute to reducing the social dis-
tance between commercial housing owners and relocated resi-
dents? To achieve this objective, this study divides social capital
into ofine and online social capital according to the charac-
teristics of “presence and absence of social interaction,” using
sampled data from the “China Urban Resident Living Space
Survey” (2019). Specically, it attempts to analyze the effects of
online and ofine social capital on reducing the social distance
between commercial housing owners and relocated residents.
2
|
Theory and Research Hypotheses
German sociologist Simmel (1991) argued that social distance
represents the extent to which individuals subjectively perceive
they should be close to others, indicating that social distance
reects people's psychological attitudes (K. Wang and Liu 2015).
This pertains to the overall perception and acceptance of one
group of actors toward another. Based on existing research
ndings, two theoretical perspectives can be identied to study
the factors inuencing social distance.
First, the status structure view considers social structure from
the perspective of individual group membership and status
characteristics. Social actors naturally segregate into different
groups based on their awareness of class distinctions. Social
distance between classes stems from differences in objective
status attributes, leading to signicant disparities in lifestyle,
cultural orientation, and values. Attributes such as socioeco-
nomic status and demographic characteristics contribute to this
perception of status. Numerous studies have shown that income
disparities affect the frequency of interaction, willingness to
engage, and the emergence of biases between different groups,
directly inuencing the deepening or mitigation of social dis-
tance (G. Lu 2006; Guo and Yang 2005; G. Lu 2003). The impact
of education on social distance is also signicant, as individuals
with higher education levels tend to have smaller subjective
social distances from other groups and are more willing to
accommodate and accept those who differ from the mainstream
(McCutcheon 1985). However, subjective class consciousness
has a negative effect, with individuals from higher classes un-
willing to interact with lower classes, while those from lower
classes, due to feelings of inferiority, are also unwilling to
interact with higher classes, leading to an increasing social
distance between them (Jasso and Milgrom 2004). Institutional
attributes, such as hukou status, may further entrench status
hierarchies, and farmers may be marginalized in community
decision‐making (Hsing 2010). Regarding demographic charac-
teristics, race is a signicant factor in the perception of social
distance, with some scholars directly equating social distance
with racial differences (Canon and Mathews 1971; Warner and
Dennis 1970). Additionally, gender inuences individuals' sub-
jective social distance, with women generally feeling a greater
social distance than men (G. Lu 2006). Furthermore, age has a
signicant effect on perceived social distance, with social dis-
tance exhibiting an inverted U‐shaped relationship with age (Hu
and Wang 2012). Marital status also cannot be overlooked in its
impact on social distance (Hipp and Perrin 2009; Fischer 1982).
The second perspective is the social network view, also known
as the social interaction perspective, which focuses on the
structural interactions within a given geographical area. In-
dividuals from different classes and groups inevitably have more
opportunities for interaction when living in the same
geographical area. In the same community, owing to collective
2 of 13 Sociology Compass, 2025

interests, public affairs, and organizational activities, structural
contacts may increase social interaction between different
classes, thereby narrowing social distances. This type of struc-
tural social interaction within the community falls under the
“macro social capital.” Bian and Lei (2017) argue that social
capital, which includes norms of reciprocity and trust arising
from social relationships, can enhance social efciency by
facilitating coordinated actions. Their analysis focused on how
specic groups develop and maintain social capital as a collec-
tive asset and how these social assets affect the life opportunities
of their members (Zhang 2003). Among all forms of social
capital, bridging and bonding social capital, or inclusive and
exclusive social capital, are perhaps the most important (Put-
nam 2011). Depending on the nature of social networks, reci-
procity, and trust, social capital is typically divided into these
two categories. The former connects people with different ages,
races, occupations, and other characteristics, whereas the latter
connects people with similar demographic characteristics and
social classes (Putnam and Sun 2019). Bridging social capital
represents diverse and inclusive social relationships, also known
as heterogeneous social interactions, while bonding social cap-
ital aids actors in establishing specic norms to ensure the
uniformity and unity of network members, also known as ho-
mogeneous social interactions (Putnam 2011).
Higher levels of social capital can lead to smaller social dis-
tances between different social classes. G. Wang and Wu (2011)
found that as the quality of social capital among urban residents
increases, the likelihood of including migrant workers in their
social networks also increases. This greater understanding of the
migrant worker community reduces prejudice and increases the
willingness to befriend them. For migrant workers, social cap-
ital not only serves these functions but also helps them change
their self‐identity and further integrate into the city. Moreover,
higher diversity in social networks strengthens the effect of
reducing social distance (Hu and Wang 2012). A person's social
circle is more likely to be exclusionary when homogeneous,
whereas it is more inclusive if it is heterogeneous. Pan and
Zou (2017) found that intra‐group interactions widen the social
distance between migrants and urban residents, while inter‐
group interactions narrow the social distance between them.
The social network perspective builds upon international theo-
retical foundations of social capital and urban interaction. Put-
nam's (2011) seminal distinction between bridging and bonding
social capital provides a critical framework: while bonding
capital reinforces homogeneity within groups (e.g., ethnic en-
claves), bridging capital facilitates connections across social di-
vides through weak ties (Granovetter 1973). Sampson's (2012)
research in Chicago demonstrated how neighborhood‐level
collective efcacy—a form of ofine bridging capital—
mediates racial boundaries via shared civic engagement.
Comparative studies further reveal contextual variations. Blok-
land and Nast's (2014) analysis of Berlin's mixed‐housing pro-
jects revealed that physical proximity alone cannot reduce
symbolic boundaries unless supported by institutionalized
interaction platforms. In the digital realm, Ellison et al. (2014)
proposed the social compensation hypothesis, arguing that on-
line interactions compensate for weak ofine ties in spatially
segregated communities. These international insights establish
social capital's dual role as both connector and divider,
contingent on its typology, medium, and policy context. The
Chinese case extends international theories. Unlike Western
suburban‐urban divides (Massey and Denton 1993), China's
hybrid communities—co‐locating commercial and resettlement
housing—create institutionalized spatial adjacency where class
boundaries are simultaneously blurred by shared infrastructure
yet reinforced by property rights hierarchies.
In summary, viewing social distance between different groups
from a status structure perspective is a static cognition, resulting
in limited understanding between different actors. By contrast,
the social network perspective emphasizes the contact between
different groups during the interaction process, resulting in a
more comprehensive understanding between different actors.
This is a dynamic process that has the potential to promote
social distance between different classes and groups. Therefore,
this study aims to identify factors that reduce social distance
between resettled households and commercial property owners,
emphasizing the dynamic and changing perspective of the social
network view (social interaction perspective) and focusing on
the inuence of macro social capital on social distance between
urban community commercial property owners and resettled
households.
Previous studies have offered a wealth of empirical evidence and
theoretical reections; however, there are notable gaps that this
paper seeks to address. First, in terms of measurement, some
studies have conceptualized collective social capital following
Putnam's framework but operationalized it through measures of
individual social capital. This approach has led to confusion
between different levels of social capital concepts. Meanwhile,
this study establishes that the unit of analysis remains at the
group level, aiming to avoid category confusion as much as
possible. Second, many studies have not considered the impact
of online social capital on social distance. With the deep pene-
tration of the internet into people's lives, internet‐based social
capital inevitably becomes an indispensable component of social
capital, playing an irreplaceable role in community life. There-
fore, this study includes online social capital as an explanatory
variable. Third, existing studies are relatively uniform, with
limited research on social distance among urban residents of
different social strata. Hence, further exploration and rene-
ment of the characteristics of urban living spaces are needed.
Our study focuses on a typical case of resettlement housing. In
the context of urban development in China, where a plot of land
slated for development already contains properties, the owners
can choose cash compensation or resettlement, that is, without
receiving cash compensation. However, the developer must
provide a new property that meets certain standards. These
properties are known as resettlement housing, and their owners
are resettlement households. Many developers choose to build
resettlement and commercial housing within the same com-
munity; however, the social strata of commercial housing
buyers are generally higher than those of resettlement house-
holds, especially in rst‐tier cities. Commercial housing buyers
often have strong economic capabilities and decent jobs,
whereas resettlement households are typically farmers or low‐
income citizens, leading to systematic differences in educa-
tion, lifestyle, and values. Additionally, owing to different
methods of acquiring property and types of property rights,
residents within the same community may experience various
3 of 13

conicts of interest and conceptual conicts. Commercial
housing owners worry about the uncivilized behaviors of
resettlement households and are reluctant to share community
facilities (e.g., gardens and tness equipment) with them, hav-
ing paid for these amenities. Resettlement households, whose
homes have been transformed, may be treated differently in new
communities, and their legitimate rights may not be guaranteed.
The actual conicts of interest between the two can be alleviated
by regulating the behavior of developers, but the prejudices that
lead to social distancing require the strength of the community
to resolve.
Based on the above literature review, the key to reducing social
distance through social capital lies in deepening understanding
between different groups when members of other groups are
present in an actor's network, thereby eliminating previous
prejudices and increasing the willingness for closeness. Actors
with more bridging and heterogeneous social capital have
smaller social distances to other groups, while those with more
bonding and homogeneous social capital have greater social
distances to other groups.
The ofine community social capital of community residents is
mostly formed through participation in community activities or
community life. When participating in ofine community ac-
tivities, residents cannot choose all participants according to
their preferences, as they cannot fully know who will participate
before the event and cannot intervene in the participant list.
Therefore, it is difcult for them to avoid or exclude members of
other property groups. Furthermore, since they live in the same
community, actors objectively cannot completely avoid mem-
bers of different groups in daily life, and the convenience of
contact increases the possibility of interaction. Moreover, living
in the same place also leads to joint participation in rights
protection activities due to common interests, such as opposing
the construction of garbage incineration plants or chemical
plants nearby, which can alleviate internal community conicts.
Identication with each other increases during the handling of
common affairs, thereby reducing the social distance between
them. In summary, at least three aspects of ofine social
interaction within the community can serve as a bridge con-
necting commercial housing owners and resettlement house-
holds, cultivating bridging social capital between the two
property groups. Hence, misunderstandings will gradually
diminish or weaken after both sides establish relationships
through community interactions, thus narrowing the social
distance between them. Based on this, this study proposes the
following hypothesis.
Hypothesis 1. An increase in ofine social capital corresponds
to a reduction in the social distance between commercial housing
owners and resettlement households.
Although ofine social capital plays a signicant role in
enhancing the social integration of different property groups
within communities, its effectiveness is constrained by charac-
teristics such as high mobility and heterogeneity in modern
urban societies. Middle‐aged and young residents of urban
communities spend most of their time outside their neighbor-
hoods, venturing out during the day to work in various parts of
the city, which leaves them with limited opportunities to
participate in ofine community activities or advocacy actions.
Thus, the impact of ofine social capital is more pronounced
among the elderly and children, with limited inuence on the
middle‐aged and young residents. In this context, online social
capital emerges as a vital form of social capital, showcasing its
superiority. First, the virtual space transcends the limitations of
time and space, enabling middle‐aged and young people to
establish connections with their neighbors through online me-
diums such as community WeChat groups, forums, and public
accounts, thereby enhancing communication and fostering so-
cial capital. Second, online platforms provide access to a wider
range of acquaintances; a typical community activity might be
limited to a dozen participants due to space constraints, whereas
a community WeChat group can facilitate connections among
hundreds of individuals, increasing the likelihood of actors
encountering members of other property groups. Therefore,
online social capital complements ofine social capital in at
least two respects, nurturing bridging social capital, or hetero-
geneous social capital, between commercial housing owners and
resettlement households, and consequently narrowing the social
distance between them. Based on this observation, this study
proposes the following hypothesis.
Hypothesis 2. Increased online social capital contributes to a
reduction in the social distance between commercial housing
owners and resettlement households.
3
|
Materials and Methods
3.1
|
Data Sources
In this study, the data were obtained from the “Living Space
Survey of Chinese urban residents,” which was carried out by
the Research Center of Sociological Theory and Method of
Renmin University of China. Conducted by Professor Shaojie
Liu, this survey adopted the phased systematic sampling
method, combined with quota and random sampling (Zhou and
Yiqi 2021). It was carried out in 17 large and medium‐sized
cities, including Beijing, Shanghai, Xi'an, Guangzhou,
Guiyang, Changchun, Hefei, and Nanchang from 2018 to 2019.
The cities surveyed cover a representative cross‐section of major
regions in China. The survey targeted urban residents aged 18
and above, including two different property rights groups:
commercial housing owners and resettlement residents. The
main survey included respondents' housing conditions, com-
munity life, overall life experience, and Internet use. After data
cleaning, we obtained 1411 valid samples, including 776 com-
mercial housing owners and 635 resettlement residents.
3.2
|
Variable Measurement
3.2.1
|
Dependent Variable: Social Distance
Social distance is represented by the degree of willingness of one
property group to approach the other group within the resi-
dential community. Based on the Bogardus Social Distance
Scale (Bogardus 1925) and the actual situation, this study con-
structed a scale comprising ve questions. Specically, the
4 of 13 Sociology Compass, 2025

questions are: “Are you willing to do the following activities
with resettled residents/home buyers: ①live in the same resi-
dential area; ②share public resources in the community; ③
participate in community activities together; ④engage in com-
munity management affairs together; and ⑤become friends.”
These ve questions are progressively ranked in terms of in-
timacy, with assigned values increasing from 1 to 5. The re-
sponses indicating “willing” are assigned a value of 3,
“indifferent” a value of 2, and “unwilling” a value of 1. The
product of the score for each question and the corresponding
response value is summed to obtain the respondent's willing-
ness to approach the other group, resulting in an interval vari-
able named “social distance.” A higher score indicates a smaller
social distance, whereas a lower score indicates a larger social
distance.
The multiplicative scoring (the value of question �the value of
response) reects that social distance reduction is a non‐additive
process, where acceptance of high‐intimacy interactions exerts a
disproportionately greater effect than acceptance of low‐
intimacy ones. Mathematically, assigning weights (1–5) to
different interaction types reects their ordinal hierarchy in
trust thresholds, while response values (1–3) quantify willing-
ness intensity. Their product generates a weighted interaction
term that captures two statistical realities: (1) the marginal
utility of accepting higher‐intimacy interactions diminishes
nonlinearly, as trust‐building requires escalating commitment;
(2) rejection of critical thresholds creates categorical barriers
that cannot be linearly offset by concessions on trivial in-
teractions. For instance, a “willing” response to friendship
(5 �3=15) contributes quadratically more to distance reduc-
tion than “willingness” to live in the same residential area
(1�3=3). The multiplicative design thus statistically oper-
ationalizes social distance as a cost‐weighted path across the
intimacy threshold. Continuous values reect nuanced differ-
entiation in selective openness.
Table 1presents descriptive statistics for the variables in this
study. As shown in the table, the average willingness of com-
modity house owners to approach resettled residents is 32.00,
while the average willingness of resettled residents to approach
commodity house owners is signicantly higher (35.43), high-
lighting the asymmetry in social distance. However, overall,
neither resettled residents nor commodity house owners show
overt hostility toward each other.
3.2.2
|
Independent Variable: Social Capital
Ofine social capital is represented by the degree of residents'
participation in community social activities and reciprocity.
Referring to the method proposed by Gui and Huang (2008), a
new continuous variable named “ofine social capital” is con-
structed using eleven four‐point ordinal variables indicating
willingness level. These 11 variables measure residents' perfor-
mance in ve dimensions: public affairs participation, neigh-
borhood interaction, voluntarism, cohesion, and social support
within the community. Specic questions are shown in Table 2.
The answers to the 11 questions in the table are assigned values
TABLE 1 |Descriptive statistics of variables.
Commercial housing
(N=776)
Resettled residents
(N=635)
Signicance of mean difference
(p‐value)Mean/%
Standard
deviation Mean/%
Standard
deviation
Social distance 31.991 9.437 35.431 9.171 0.0000
Ofine social capital 22.706 5.996 22.706 6.045 0.9983
Online social capital 0.796 0.877 0.724 0.826 0.1156
Real 56.44 56.69 0.925
Age 44.541 14.141 47.337 14.081 0.911
Ethnicity (minorities) 3.61 4.57 0.363
Marital status (not in
marriage)
13.66 15.91 0.236
Education 0.000
High school or Vocational
school
42.65 50.39
Bachelor's degree or above 40.98 13.54
Personal annual income
(yuan)
73,025.856 80,284.19 41,701.439 47,973.766 0.000
Household registration (not
local)
17.78 3.94 0.000
Urban and rural (rural) 36.98 47.40 0.000
Subjective class 2.648 0.716 2.339 0.776 0.031
5 of 13

from 1 to four in increasing order of degree. Then, the scores are
summed to obtain the variable “ofine social capital.” A higher
score indicates more ofine social capital, whereas a lower score
indicates less ofine social capital. From Table 1, the mean
value of ofine social capital for commodity house owners is
consistent with that of resettled residents, both being 22.71,
indicating that there is no signicant difference in ofine social
capital between the two groups, which are both at a medium
level.
Online social capital is measured as a summative index (0–3)
reecting engagement across three platforms: (1) membership
in community WeChat groups, (2) following community public
accounts, and (3) registration on community forums. Each
platform participation is coded as 1 (present) or 0 (absent), with
the total score representing the diversity of online social ties.
This approach captures the multidimensional nature of online
social capital (Lin 2001), distinguishes between minimal and
maximum digital engagement, and aligns methodologically with
our measurement of the ofine social capital, thereby
enhancing analytical consistency. As shown in Table 1, both
commercial homeowners (M=0.796, SD =0.877) and resettled
residents (M=0.724, SD =0.826) exhibit similarly low levels of
online social capital, with no statistically signicant difference
between groups (F=2.48, p=0.1156). Nearly half of the
samples have no online social capital, indicating limited digital
community participation overall.
3.2.3
|
Control Variables
Personal Annual Income. From Table 1, the mean personal
annual income of commodity house owners is 73,025.86 yuan,
which is signicantly higher than the 41,701.44 yuan of resettled
residents. Personal annual income exhibits a strong skewed
distribution. Therefore, the logarithm of personal annual in-
come is included in our analysis model.
Education Level. It is an ordinal variable with progressive
assignment from low to high, divided into three levels.
Compulsory education and below are assigned a value of 1, high
school and junior college are assigned a value of 2, and bache-
lor's degree and above are assigned a value of 3. As shown in
Table 1, the overall education level of resettled residents is
signicantly lower than that of commodity house owners.
Subjective Social Class. It is measured using a ve‐point Likert
scale with assignment from low to high. The mean value for
commodity house owners is 2.65, signicantly higher than that
of resettled residents (2.34).
Through these three variables—income, education level, and
subjective social class—which reect socioeconomic status, we
can initially conclude that commodity house owners and
resettled residents belong to two different social classes, with
commodity house owners occupying a higher social class than
TABLE 2 |Ofine social capital measurement question Set.
Dimension Concrete problems
Participation in public affairs The frequency of your actions in the past year is: Never; Occasionally;
Sometimes; often
Discussing issues or events in the community with neighbors;
Participate in activities organized by the community neighborhood committee
or property management department;
Participate in community meetings;
Participate in community organizations.
Neighborhood interaction You and your neighbors: Do not know each other; I know who they are, but I
haven't spoken with them yet; mutual understanding and occasional
communication; I am quite familiar and frequently interact with others.
The frequency of your actions in the past year is: Never; Occasionally;
Sometimes; often
Say hello to neighbors when meeting.
Voluntarism The frequency of your actions in the past year is: Never; Occasionally;
Sometimes; often
Reect on community management issues.
Cohesion How do you usually pass the time: Never; Occasionally; Sometimes; often
Participate in cultural and sports activities within the community;
Attend training classes or salons within the community.
Social support The frequency of your actions in the past year is: Never; Occasionally;
Sometimes; often
Ask neighbors for help;
Lend or give tools, food, or other items to neighbors.
6 of 13 Sociology Compass, 2025

resettled residents. The preliminary presupposition of this study
is consistent with the data.
Household Registration. Non‐local residence permit is assigned a
value of 0, while local residence permit is assigned a value of 1. As
shown in Table 1, the proportion of commodity house owners
with local residence permits is 82.22%, signicantly lower than
that of resettled residents (96.06%). Since relocation is generally a
policy benet exclusive to locals, this result is logical.
Urban and Rural. Rural household registration is assigned a
value of 0, whereas urban household registration is assigned a
value of 1. As shown in Table 1, most commodity house owners
and resettled residents hold urban household registrations.
Among them, the proportion of commodity house owners with
urban household registrations is 63.02%, signicantly higher
than that of resettled residents (52.60%).
The partial differences between commodity house owners and
resettled residents in terms of local and urban‐rural household
registration are related to the urban‐rural dual structure of
Chinese society and the characteristics of relocation policies.
Resettled residents are generally original villagers who have
migrated to urban areas during urban expansion. Therefore,
compared with commodity house owners, they are mostly local
rural populations.
Gender. It is a dummy variable with males assigned a value of
0 and females assigned a value of 1. As shown in Table 1, the
proportion of females is slightly higher among both commodity
house owners and resettled residents; however, there is no sig-
nicant difference between the two groups.
Age. It is an interval variable obtained by subtracting the birth
year from 2019. As shown in Table 1, there is no signicant
difference in the average age between the two groups, with
commodity house owners having an average age of 44.54 years,
slightly lower than that of resettled residents (47.34 years).
Marital Status. It is a dummy variable. Those who are not married
(including unmarried, widowed, and divorced) are assigned a
value of 0, whereas married people are assigned a value of 1. As
shown in Table 1, most respondents are married, with no sig-
nicant difference in marital status between the two groups.
Ethnicity. It is a dummy variable with Han ethnicity assigned a
value of 1 and other ethnicities (minorities) assigned a value of
0. There is no signicant difference between commodity house
owners and resettled residents in terms of ethnicity.
Gender, age, marital status, and ethnicity, as basic demographic
characteristics, show no signicant differences in their distribu-
tion among commodity house owners and resettled residents,
facilitating the comparison of model results between the two
samples.
3.3
|
Models
This study employs multiple linear regression and random for-
est models to separately analyze samples of commercial housing
owners and resettled residents. The dependent variables are
categorized into two types: the social distance of commercial
housing owners toward resettled residents and the social dis-
tance of resettled residents toward commercial housing owners.
The study analyzes the relationship between both online and
ofine social capital and the social distance between these two
property rights groups.
The relationship between social capital and social distance is not
always linear, and further modeling to consider the nonlinear
relationship between variables is necessary. Simultaneously,
there may be collinearity between online social capital and
ofine social capital. The random forest model can analyze the
nonlinear relationship between dependent and independent
variables and solve collinearity problems.
The random forest model is an ensemble learning algorithm
based on decision trees. It improves prediction accuracy and
stability by averaging the prediction results from multiple de-
cision trees. Each decision tree is built based on a randomly
selected subset of training samples and a randomly selected
subset of features. This dual randomness gives the model strong
generalization ability and resistance to overtting.
The relationship between social capital and social distance may
constitute a complex system involving multiple interactions.
Multiple linear regression assumes a linear relationship be-
tween the dependent and independent variables, whereas the
random forest model is not limited by this linear assumption
and can capture such complex nonlinear relationships. Second,
multiple linear regression is highly sensitive to multicollinearity
among features. In the case of a high correlation between in-
dependent variables, the stability and interpretability of the
regression model may be affected. However, the random forest
model effectively reduces the impact of feature correlations by
randomly selecting feature subsets to construct each tree,
enhancing the model's robustness. Finally, the random forest
algorithm can calculate the contribution of each feature to the
prediction results, automatically assessing feature importance.
This helps researchers identify key factors affecting social dis-
tance without requiring tedious feature selection and ltering
processes beforehand, further clarifying the relative importance
of online and ofine social capital.
4
|
Results
4.1
|
Multiple Linear Models
Initially, this paper utilizes multiple linear regression to
examine the impact of the two core independent variables on
the social distance between the two property rights groups,
resulting in Models 1 to 4. Subsequently, both online and ofine
social capital are incorporated into the models simultaneously,
yielding Models 5 and 6. The results are presented in Table 3.
Based on the results of Models 1 and 2, residents with richer
ofine social capital, whether they are commercial housing
owners or resettled residents, show signicantly higher will-
ingness to be close to members of the other property rights
7 of 13

group, resulting in a smaller social distance between the two
groups, thus conrming Hypothesis 1. Similarly, according to
the results of Models 3 and 4, residents with more abundant
online social capital, regardless of being commercial housing
owners or resettled residents, demonstrate a notably increased
closeness toward members of the other property rights group,
leading to a reduced social distance, thereby verifying Hypoth-
esis 2. However, upon comparing Models 1 and 3, as well as
Models 2 and 4, it is observed that the signicance of the
regression coefcient for online social capital is lower than that
for ofine social capital.
The results of Models 5 and six indicate that when controlling
for online social capital, ofine social capital still signicantly
reduces social distance. However, when controlling for ofine
social capital, online social capital does not signicantly reduce
social distance. This suggests that ofine social capital demon-
strates stronger stability in bridging the social distance between
different property rights groups. Participating in ofine com-
munity activities, rights protection actions, neighborhood
mutual assistance, and other similar events is thus more effec-
tive in promoting integration among different social classes.
Although the internet has the potential to improve neighbor-
hood relations, ofine social capital plays a more crucial role,
even serving as a key factor with a fundamental impact on
resident interactions. Online social capital does not replace but
rather complements the role of ofine social capital in
improving social distance within urban communities.
TABLE 3 |Impact of online and ofine social capital on the social distance.
Ofine social capital Online social capital Full Model
Commercial
(Model 1)
Resettled
(Model 2)
Commercial
(Model 3)
Resettled
(Model 4)
Commercial
(Model 5)
Resettled
(Model 6)
Ofine social capital 0.215*** 0.248*** 0.177** 0.237***
(0.0569) (0.0625) (0.0623) (0.0709)
Online social capital 1.118** 0.975* 0.631 0.171
(0.386) (0.452) (0.420) (0.509)
Gender 1.650* 0.230 1.612* 0.0513 1.678* 0.240
(Control group: Female) (0.690) (0.751) (0.692) (0.757) (0.689) (0.753)
Age −0.0548* −0.0372 −0.0325 −0.00920 −0.0489 −0.0359
(0.0267) (0.0309) (0.0265) (0.0304) (0.0270) (0.0312)
Ethnicity 3.432 0.365 3.841* 0.572 3.573* 0.401
(Control group: Minorities) (1.804) (1.744) (1.811) (1.762) (1.805) (1.748)
Marital status 0.676 −0.133 0.584 −0.143 0.642 −0.162
(Control group: not in marriage) (1.010) (1.032) (1.014) (1.045) (1.009) (1.036)
Urban and rural −0.388 1.249 −0.437 1.160 −0.395 1.278
(Control group: rural) (0.726) (0.781) (0.729) (0.791) (0.726) (0.786)
Household registration −1.786 1.338 −1.745 1.027 −1.799 1.296
(Control group: not local) (0.923) (1.887) (0.926) (1.907) (0.922) (1.893)
High school and Vocational
school
−0.692 −1.581 −0.662 −1.484 −0.693 −1.577
(Control group: Compulsory
education or below)
(1.015) (0.862) (1.019) (0.869) (1.014) (0.863)
Bachelor's degree or above −2.148 −0.178 −2.298* 0.214 −2.241* −0.195
(Control group: Compulsory
education or below)
(1.112) (1.360) (1.117) (1.367) (1.112) (1.362)
Annual income logarithm −0.0183 0.0232 −0.0176 0.0576 −0.0248 0.0250
(0.124) (0.110) (0.124) (0.110) (0.124) (0.110)
Subjective class −0.728 −0.552 −0.564 −0.438 −0.695 −0.553
(0.478) (0.472) (0.478) (0.475) (0.478) (0.472)
Intercept 29.94*** 31.18*** 32.25*** 34.33*** 29.95*** 31.25***
(2.896) (3.421) (2.792) (3.330) (2.894) (3.430)
Sample size 776 635 776 635 776 635
Adjusted R
2
0.0347 0.0195 0.0274 0.0021 0.0363 0.0181
Note: *p<0.05, **p<0.01, ***p<0.001.
8 of 13 Sociology Compass, 2025

The signicance of the control variables remains largely
consistent across all six models, indicating good stability of the
results. Among commercial housing owners, males demonstrate
a higher willingness to be close to resettled residents compared
with females. However, among resettled residents, gender has
no signicant effect on social distance. The nding that females
tend to maintain a greater social distance aligns with previous
research. From a social role perspective, women are often so-
cially expected to exhibit more delicate and conservative traits,
which may lead them to maintain an inherent caution and
distance when facing groups with different backgrounds.
Meanwhile, based on social identity theory, women tend to
establish close relationships with individuals or groups similar
to themselves and may feel excluded from groups with signi-
cant differences, thus maintaining a greater social distance.
Additionally, age and ethnicity somewhat inuence the social
distance of commercial housing owners. When controlling for
ofine social capital, older commercial housing owners tend to
maintain a greater social distance. When controlling for online
social capital, Han Chinese commercial housing owners main-
tain a smaller social distance compared with ethnic minorities.
This may stem from the xed social networks and community
perceptions formed by older individuals over their long social
lives, leading them to adopt a conservative attitude toward
newly joined groups with different backgrounds. Meanwhile,
the mainstream cultural background of the Han ethnic group
and the diversity of online information may facilitate their
acceptance and understanding of different groups. The unex-
pected negative association between higher education (bache-
lor's degree or above) and reduced social distance among
commercial homeowners—only signicant when controlling for
online social capital—could be attributed to the fact that
educational attainment correlates with critical awareness of
property rights hierarchies, fostering strategic avoidance of
resettled groups perceived as threatening asset values. This
pattern is absent among resettled residents presumably due to
their limited educational heterogeneity and shared identities,
which override individual status markers.
By contrast, none of the control variables are signicant in the
models for the resettled residents. This may be because their
shared experience of urban renewal and demolition has a
greater impact on their identity and cognition, overshadowing
the inuence of demographic characteristics. Other control
variables, besides gender, age, education, and ethnicity, do not
signicantly affect social distance in any of the three models for
both groups. After controlling for demographic characteristics
and social capital, social class and urban‐rural residency status
no longer have a signicant impact on social distance.
To assess potential multicollinearity, we calculate variance
ination factors (VIFs) for all predictors. As shown in Table 4,
mean VIF values range from 1.19 to 1.36 across models, with
individual VIFs for key variables—online social capital
(1.02–1.36) and ofine social capital (1.05‐1.41)—well below
the critical threshold of 5. This conrms that collinearity be-
tween predictors does not substantially distort parameter
estimates.
4.2
|
Random Forest Models
While the results of the multiple linear regression models are
satisfactory and the linear regression models show no severe
multicollinearity (all of the mean VIFs <2), we further employ
random forest models to address the limitation of potential
nonlinear relationships between social capital and distance
reduction. By randomly subsampling features during tree con-
struction (Breiman 2001), random forests inherently mitigate
correlated predictors' instability, which provides robustness
checks against linear model assumptions, captures potential
nonlinear effects, and further reduces collinearity risk.
No missing values are found in the commercial housing owner
and relocated household samples. The independent variables in
the model include ofine social capital, online social capital,
TABLE 4 |Values of the variance ination factor.
Model 1 Model 2 Model 3 Model 4 Model 5 Model 6
Ofine social capital 1.05 1.10 1.26 1.41
Online social capital 1.02 1.06 1.23 1.36
Gender 1.06 1.07 1.06 1.06 1.06 1.07
Age 1.29 1.46 1.26 1.38 1.32 1.48
Ethnicity 1.02 1.02 1.02 1.02 1.02 1.02
Marital status 1.09 1.10 1.09 1.10 1.09 1.10
Urban and rural 1.11 1.17 1.11 1.18 1.11 1.18
Household registration 1.12 1.04 1.12 1.04 1.12 1.04
High school and Vocational school 2.27 1.43 2.27 1.43 2.27 1.43
Bachelor's degree or above 2.70 1.67 2.71 1.66 2.71 1.67
Annual income logarithm 1.11 1.12 1.11 1.11 1.11 1.12
Subjective class 1.05 1.03 1.05 1.03 1.06 1.03
Mean VIF 1.35 1.20 1.35 1.19 1.36 1.24
9 of 13

gender, age, ethnicity, marriage, education, social class, per-
sonal annual income, urban‐rural residency, and household
registration. The dependent variable is social distance. Modeling
is performed separately for the commercial housing owner and
relocated household samples, with each model consisting of 500
trees. Table 5shows the importance values of each independent
variable's impact on social distance for the two groups.
Our random forest models reveal critical nuances in how social
capital and demographic factors shape intergroup social dis-
tance. IncMSE represents the percentage increase in the model's
mean squared error when a variable is randomly permuted,
measuring the importance of the variable to the prediction ac-
curacy of the model. A higher value indicates that the variable is
more important for the prediction accuracy of the model. For
commercial homeowners, the IncMSE for ofine social capital
is 7.56%, meaning that if the values of ofine social capital are
randomly permuted, the mean squared error of the model will
increase by approximately 7.56%. The IncMSE for online social
capital is 4.57%, which is lower than that of ofine social capital.
IncNodePurity is calculated based on the reduction in impurity
at each node in the decision tree. It represents the sum of the
improvements in node purity when a variable is used for split-
ting during tree construction. A higher node purity indicates a
greater contribution of the variable to improving node purity
during tree construction, making it more valuable for classi-
cation or regression tasks. In the commercial housing sample,
the IncNodePurity for ofine social capital is 10,500.62, indi-
cating that the total improvement in node purity brought about
by ofine social capital during splits is 10,500.62 across all trees.
The IncNodePurity for online social capital is 3729.29, which is
lower than that of ofine social capital. The importance values
of ofine social capital are high in all variables, while those of
the online social capital are in the middle. This conclusion also
holds true for resettled residents. These results corroborate the
linear regression ndings that face‐to‐face interactions are more
effective in reducing social distance and signicantly bridge
class divides, suggesting that digital engagement supplements
rather than substitutes ofine ties—a pattern consistent across
both property groups.
Notably, the results of IncMSE and IncNodePurity are not al-
ways consistent. A variable may perform well in improving node
purity but have a relatively small impact on the mean squared
error of the model. This depends on the characteristics of the
data and the specic application of the model. To demonstrate
the importance of each independent variable to the prediction
accuracy of the model, a scatter plot is created for each inde-
pendent variable's importance in the models for commercial
housing owners and resettled residents. The height of each point
represents the importance score of the corresponding variable,
and the variables are sorted in descending order of importance,
as shown in Figure 1.
For commercial homeowners, gender signicantly predicts
larger social distance (IncMSE =9.06%, highest among all var-
iables), resonating with social role theory that women prioritize
in‐group safety in class‐heterogeneous settings (Eagly 1987).
Meanwhile, urban‐rural divides disproportionately affect reset-
tled residents (IncMSE =9.91% vs. 2.34% for commercial
owners), which still shows that people in resettlement housing
amplify their reliance on urban‐rural identity as a boundary
marker (Hsing 2010).
Notably, ethnicity negatively impacts the prediction in these two
models (IncMSE = −0.62% and −1.15%), implying that
permuting ethnicity values reduces model error. This counter-
intuitive result aligns with China's ethnic situation. As 95.96% of
our sample are Han Chinese, this creates minimal ethnic het-
erogeneity to exploit for prediction. The model's attempt to
overt this quasi‐homogeneous distribution introduces noise,
thereby inating error when ethnicity is preserved—a phe-
nomenon observed in low‐diversity contexts.
The difference between IncMSE and IncNodePurity highlights
some context‐dependent variable roles. For instance, age shows
TABLE 5 |Results of the random forest models.
Owners of commercial housing Resettled households
IncMSE(%) IncNodePurity IncMSE(%) IncNodePurity
Ofine social capital 7.56 10,500.62 13.68 8901.37
Online social capital 4.57 3729.29 7.28 2897.42
Gender 9.06 2310.15 3.31 1609.37
Age 6.40 12,380.13 7.68 9472.79
Ethnicity −0.62 914.81 −1.15 837.57
Urban and rural 2.34 1894.34 9.91 1835.40
Household registration 0.83 1444.84 1.84 700.86
Marital status 1.74 1287.38 2.97 1180.09
Education 7.35 3064.52 9.10 2512.02
Personal annual income 5.46 10,458.50 7.13 7689.79
Subjective class 3.42 3947.67 2.03 3220.66
Sample size 776 635
Correlation coefcient 0.94 0.94
10 of 13 Sociology Compass, 2025

moderate IncMSE in the two models (6.40% and 7.68%) but high
IncNodePurity (12,380.13 and 9472.79), indicating that while
age weakly predicts distance magnitude, it critically structures
decision‐tree splits—likely reecting lifecycle variations in
openness (older residents resist cross‐class ties due to
entrenched networks).
To evaluate the overall prediction effectiveness of the model, the
correlation coefcient between the predictions and the actual
observed values was calculated. The results are shown in the last
column of Table 5. The correlation coefcients for both the
commercial housing owners' and resettled residents' models are
0.94, indicating a strong positive correlation between the pre-
dicted and actual observed values, and this relationship is linear.
This means that our prediction models are able to neatly capture
the trend of actual data changes and complex social dynamics.
5
|
Discussion and Conclusion
Utilizing data from the 2019 Survey on Living Spaces of Urban
Residents in China, this study employed multiple linear
regression and random forest modeling to investigate the in-
uence of online and ofine social capital on the social distance
between two property right groups: commercial housing owners
and resettled residents. The research ndings are as follows:
First, ofine social capital plays a signicant and positive role in
enhancing social proximity among diverse property right
groups, serving as a fundamental function. Increased social
interaction among both commercial housing owners and
resettled residents contributes to bridging their social distances.
This discovery aligns with the social capital theory advocated by
scholars like Putnam, underscoring the importance of actual
community interactions. In the rapidly urbanizing context of
China, the proximity of living spaces between commercial
housing owners and resettled residents offers more opportu-
nities for interaction. Such face‐to‐face exchanges facilitate
mutual understanding, reduce misunderstandings, and ulti-
mately narrow the social distance between the two groups. This
conclusion also reinforces the social network perspective that
“structural social interaction aids in mutual understanding.”
Second, online social capital can also diminish the social dis-
tance between the two resident groups, albeit with a slightly
weaker effect compared with ofine social capital. Its role
primarily manifests as a complementary effect to ofine social
capital. This outcome reects the social characteristics of the
internet era: while online communication offers the advantage
of transcending physical space and facilitating communication
among different groups, the relationships it establishes may not
be as deep or authentic as ofine interactions due to the frag-
mented and virtual nature of online information. Nonetheless,
the positive impact of online social capital in reducing social
distance should not be overlooked, especially in modern society
where the internet has become a crucial channel for informa-
tion acquisition and network building.
Further, this study emphasized rigorous denitions in terms of
measurement, effectively avoiding confusion between different
levels of social capital concepts. In contrast to previous studies
that have primarily focused on the impact of ofine social
capital on community life (Lin 2001), this study incorporates
online social capital into the analytical framework to respond to
the new changes in social capital forms in the internet age.
Compared with previous studies, this research focuses on the
case of relocated housing the embodies Chinese characteristics.
This not only enriches empirical research on the relationship
between social capital and social distance but also provides a
new perspective for understanding the dynamic relationships
among Chinese social classes.
In summary, this study not only veries the crucial role of social
capital in reducing social distance, but the ndings also align
with existing research suggesting that social capital can elimi-
nate prejudice and increase the willingness of both parties to be
close. Furthermore, it distinguishes between the different effects
of online and ofine social capital. These discoveries hold sig-
nicant academic and practical value for comprehending social
conicts and exploring effective ways for community integra-
tion during the urbanization process in contemporary China.
Despite its contributions, this study has certain limitations.
First, it does not differentiate between various types of com-
munity activities. Future research can further explore how
different types of community activities inuence the formation
and accumulation of social capital and how these activities
affect the process of reducing social distance. Second, this study
primarily focuses on the role of social capital in narrowing social
distance without comprehensively considering other possible
inuencing factors, such as policy environment and community
facilities. Future research can further expand the analytical
FIGURE 1
|Variable Prediction Importance Ranking Chart for Random Forest Models. U&R represents urban and rural, while HR represents
household registration.
11 of 13

framework, providing a more comprehensive understanding of
the factors and mechanisms that inuence social distance.
Third, this study mainly relies on cross‐sectional data for anal-
ysis. Introducing longitudinal data in the future can thoroughly
explore the dynamic relationship between social capital and
social distance.
Based on our ndings, we propose the following policy recom-
mendations to promote community development and narrow the
social distance among different groups within the community:
First, the organization of ofine community activities should be
strengthened to facilitate communication among residents.
Community management agencies should regularly organize
various ofine events, such as community cultural festivals,
parent‐child activities, sports competitions to encourage
different groups, including commercial housing owners and
resettled residents, to participate. This can enhance mutual
understanding, reduce prejudices, and shorten social distances.
Additionally, activities should be diversied to meet the needs
of residents of different age groups and interests, improving
resident participation and sense of belonging.
Second, online platforms should be utilized to create virtual
communities and enhance online social capital. Community
management agencies should establish and maintain commu-
nity online platforms, such as WeChat groups, ofcial accounts,
and community forums, to provide residents with a space for
online communication. Through these platforms, residents can
overcome time and space constraints. This not only enhances
the community participation of young and middle‐aged resi-
dents but also cultivates trust and cooperation among residents
through information sharing and exchange of opinions, thereby
promoting the accumulation of social capital.
Third, mixed‐income housing models should be promoted to
facilitate integration among different social classes. Urban
planning and housing policies should encourage the develop-
ment of mixed‐income housing models to avoid spatial segre-
gation of social classes. Through reasonable housing allocation
policies, residents from different social backgrounds can coexist
within the same community. Further, community management
agencies should provide equitable services and resources to
ensure that every resident can enjoy the fruits of community
development. This can help reduce social distance and promote
social integration and harmony.
Acknowledgements
We would like to thank the anonymous reviewer and the editors for the
helpful remarks and useful feedback that improved this paper.
Conicts of Interest
The authors declare no conicts of interest.
Data Availability Statement
Data available on request from the authors. The data that support the
ndings of this study are available from the corresponding author, upon
reasonable request.
References
Bian, Y., and M. Lei. 2017. “Between Virtuality and Reality: Trans-
formation of Social Capital From Virtual Space to Physical Space.”
Journal of Social Sciences of Jilin University, no. 3.
Blokland, T., and J. Nast. 2014. “Belonging in Berlin’s Mixed Neigh-
bourhoods.” International Journal of Urban and Regional Research 38,
no. 4: 1142–1159. https://doi.org/10.1111/1468‐2427.12126.
Bogardus, E. S. 1925. “Measuring Social Distance.” Journal of Applied
Sociology 9, no. 2.
Bourdieu, P. 1984. Distinction A Social Critique of the Judgement of
Taste. Routledge and Kegan Paul Limited.
Breiman, L. 2001. “Random Forests.” Machine Learning 45, no. 1: 5–32.
https://doi.org/10.1023/A:1010933404324.
Canon, L. K., and K. Mathews Jr. 1971. “Ethnicity, Belief, Social Dis-
tance and Interpersonal Evaluation: A Methodological Critique.” Soci-
ometry 34, no. 4: 515–523. https://doi.org/10.2307/2786197.
Eagly, A. H. 1987. Sex Differences in Social Behavior: A Social‐Role
Interpretation. Lawrence Erlbaum Associates.
Ellison, N. B., R. Gray, C. Lampe, and A. T. Fiore. 2014. “Social Capital
and Resource Requests on Facebook.” New Media & Society 16, no. 7:
1104–1121. https://doi.org/10.1177/1461444814543998.
Fischer, C. S. 1982. To Dwell Among Friends: Personal Networks in Town
and City. University of Chicago Press.
Granovetter, M. S. 1973. “The Strength of Weak Ties.” American Journal
of Sociology 78, no. 6: 1360–1380. https://doi.org/10.1086/225469.
Gui, Y., and R. Huang. 2008. “Measuring Community Social Capital: An
Empirical Study Based on Experiential Data.” Sociological Research,
no. 3.
Guo, X., and J. Yang. 2005. “Voluntary Segregation of Urban Migrant
Workers.” Jiangsu Administrative College Journal, no. 1.
Harvey, D. 2014. Rebel Cities: From the Right to the City to the Urban
Revolution, Translated by Ye Qimao and Ni Xiaohui. Commercial Press.
Hipp, J. R., and A. J. Perrin. 2009. “The Simultaneous Effect of Social
Distance and Physical Distance on the Formation of Neighborhood
Ties.” City & Community 8, no. 1: 5–25. https://doi.org/10.1111/j.1540‐
6040.2009.01267.x.
Hsing, Y.‐T. 2010. “Social Capital and the Social Distance of Urban
Residents From Migrant Workers.” Great Urban Transformation: Poli-
tics of Land and Property in China: Oxford, 2010; Online Edn, Oxford
Academic, 1 May 2010.
Hu, R., and X. Wang. 2012. “Social Capital and the Social Distance
Between Urban Residents and Migrant Workers From Rural Area.”
Social Science Research, no. 3: 101–106.
Jasso, G., and E. M. Milgrom. 2004. “Identity, Social Distance and Pal-
estinian Support for the Road Map. Center on Democracy.” Development
and the Rule of Law 11.
Li, Q. 2009. “Urban ‘Housing Status Groups’ During the Transition
Period.” Jiangsu Social Sciences, no. 4.
Lin, N. 2001. Social Capital: A Theory of Social Structure and Action.
Cambridge University Press. https://doi.org/10.1017/CBO978051181
5447.
Lu, G. 2003. An Empirical Study on the Social Distance between Migrant
Workers and Beijing Residents. Renmin University of China.
Lu, G. 2006. An Empirical Study on the Social Distance between Migrant
Workers and Citizens in Large Cities in China, Vol. 4. Journal of the
People’s Public Security University of China.
Lu, H. 2014. “Real Estate and Stratied Society—Reading ‘Real Estate
Class Society’.” Society, no. 4.
12 of 13 Sociology Compass, 2025

Massey, D. S., and N. A. Denton. 1993. American Apartheid: Segregation
and the Making of the Underclass. Harvard University Press.
McCutcheon, A. L. 1985. “A Latent Class Analysis of Tolerance for
Nonconformity in the American Public.” Public Opinion Quarterly 49,
no. 4: 474–488. https://doi.org/10.1086/268945.
McPherson, M., L. Smith‐Lovin, and J. M. Cook. 2001. “Birds of a
Feather: Homophily in Social Networks.” Annual Review of Sociology 27,
no. 1: 415–444. https://doi.org/10.1146/annurev.soc.27.1.415.
Pan, Z., and D. Zou. 2017. “Research on the Inuence Mechanism of
Social Distance in the Process of Urbanization of Land‐Lost Farmers –
Based on the Analysis of Social Distance Effect.” Study & Practice, no. 5.
Putnam, R. D. 1993. “The Prosperous Community: Social Capital and
Public Life.” American Prospect 13.
Putnam, R. D. 2011. Bowling Alone: The Collapse and Revival of
American Community.
Putnam, R. D., and J. Sun. 2019. “50 Years of Social Capital Research.”
Exploration & Contention, no. 3: Peking University Press.
Hu Rong, W. Xiao. “Social Science Research” (3).
Sampson, R. J. 2012. Great American City: Chicago and the Enduring
Neighborhood Effect, Vol. 1–. University of Chicago Press. https://doi.
org/10.7208/chicago/9780226834016.001.0001.
Simmel, G. 1991. The Bridge and the Door: A Collection of Essays by
Simmel. Shanghai Sanlian Bookstore.
Wang, G., and J. Wu. 2011. “Inuence Factors Analysis of Social Dis-
tance Between Migrants and Residents in Shanghai.” Sociological
Studies 25, no. 2: 28–47þ243. https://doi.org/10.19934/j.cnki.shxyj.2011.
02.002.
Wang, K., and L. Liu. 2015. “Intergroup Communication, Interpersonal
Trust, and Social Distance—A Study on the Intergroup Relationship
Between Urban Residents and Migrant Workers.” Journal of Yunnan
University (Social Sciences Edition), no. 4.
Warner, L. G., and R. M. Dennis. 1970. “Prejudice versus Discrimina-
tion: An Empirical Example and Theoretical Extension.” Social Forces
48, no. 4: 473–484. https://doi.org/10.2307/2575571.
Xu, Y., and X. Shi. 2016. “Spatial Social Stratication in Urbanization
and Chinese Mechanism.” Social Science of Beijing 11.
Zhang, W. 2003. “Social Capital: Theoretical Debates and Empirical
Research.” Sociological Research 4.
Zhou, J., and Fu Yiqi. 2021. “How Does Internet Use Affect the Com-
munity Integration of Residents: Analysis Based on Chinese Urban
Residents’ Living Space Survey.” Sociological Review of China 9, no. 5:
105–121.
13 of 13