The Impacts of Urban Environments on Community Trust of the Low-Income Group: A Case Study for the Pearl River Delta Region

Abstract

Under the burgeoning development of urbanization in China, the low-income groups have received attention recently. By applying a linear regression model and utilizing the date from the 2016 China Labor-force Dynamics Survey, this study has explored the effects of urban environments on the community trust in low-income groups, paying particular attention to the difference between local residents and migrants in the Pearl River Delta (PRD). The empirical findings suggest the following: (1) community trust in low-income groups is influenced by social environment dimension, urban space dimension, and sociodemographic characteristics. Specifically, urbanization rate, population density, POl density, land development intensity, social contact, self-rated health, and age have significant effects on the community trust of low-income groups. (2) For local residents, social environment dimension (social contact), urban space dimension (urbanization rate), and sociodemographic characteristics (political status, hukou status, age, and self-rated health) have significant effects on community trust. (3) In the case of migrants, only the sociodemographic characteristics (working in private enterprises or organizations and in agriculture) have a significant impact on community trust. According to the empirical results, the optimization of physical space and social space should consider low-income groups’ needs in livable community planning.

1. Introduction

With the increase in globalization, the economic and social structure in China has been influenced considerably [1]. At the same time, due to rapid urbanization, a large number of migrants from rural areas have moved into cities over the last several decades. Such large movements in population account for about 10% of the total population in China; Guangdong Province has the highest population growth and the permanent population at the year-end in 2020 had increased by 21.8 million since 2010 [2]. Among the cities in Guangdong Province, Guangzhou, Shenzhen, Dongguan, Foshan, and Zhuhai were the top five cities with the largest group migrants. A consequence of such mobility in the population is that the income gap among Chinese people has become more extreme [3]. Notably, the increasing mobility in the population has been responsible for the breakup of the original relationships in traditional communities which are constantly faced with the need to rebuild [4]. This is a phenomenon whereby the communities of low-income groups have suffered poor housing conditions and is manifested by the demographic heterogeneity of the community, and where lower levels of community trust have attracted attention in recent years. With reference to previous studies, the most widely accepted international poverty level is set at the average disposable personal income of the study area [5].

The hukou (household registration) system is a serious issue in China with a long history. The hukou system classifies urban migrants as temporary residents in cities, and they usually do not have a chance to change the status of hukou [6]. In addition, urban migrants are disadvantaged in terms of receiving welfare such as social and health services, especially in the case of the low-income groups [3,7]. The financial status, personal identification, resource allocation, and other factors may affect the level of community trust of the migrants, leading to a serious problem with respect to integration into the community [8]. Social exclusion occurs when low-income people are prevented from participating in social activities, which is negatively associated with the level of community trust [9]. Furthermore, the promotion of policies which advance community trust can address the negative effects of inequality on a subject’s well-being.

Notably, serious social issues can arise in the complex urban environments especially in environments where spatial inequalities related to low-income groups exist as evidenced in recent years [10]. For instance, an individual’s access to social facilities is affected by the urban space and the sociodemographic characteristics. This is so in the case of the low-income groups who require affordable housing but who suffer from having a low level of accessibility [9]. The built environment where they live can also affect the level of community trust [11,12].

At present, there are some deficiencies and gaps in the research on community trust. First, existing research has tended to focus on community trust mostly from the viewpoint of discussing social environment factors, however, the influencing factors and mechanisms of community trust have not been analyzed systematically. Second, some studies have discussed community trust mainly from the perspective of sociology and have seldom considered the impact of the elements of urban space on community trust. Third, in terms of the research object, empirical discussion of low-income groups in relation to domestic issues is insufficient, and this needs further research.

Against a background of high-quality development in China, livable community has received attention in China’s urban planning and housing policy making, which takes the requirements of low-income groups into account [13]. It is of great significance to explore the factors which influence the community trust of low-income groups and the associated mechanisms, to enhance the community trust of low-income groups and to build livable communities. This paper is organized as follows. The next section provides a theoretical framework for this study. Next, the data and variables are introduced. Empirical findings are then discussed, and the final section summarizes these findings and offers conclusions.

2. Literature Review

2.1. Trust

Trust plays a significant role in society. There is a large amount of research on trust from different disciplines, including psychology, politics, management, ethics, sociology, and economics. However, there remains some diffusion about the definition and conceptualization of trust [14,15,16,17,18]. From the perspective of social science, trust is conceived as a belief in the integrity of other individuals [19], and a collective attribute, which is applied to describe the mutual relationship among people rather than an individual’s psychological state and motivate individual to interact among residents in the community [20]. At the same time, Western researchers have long explored the links between trust and risk. Siegrist et al. found that people tend to trust governmental institutions to reduce the complexity in social life, and the way to decrease some unavoidable risk can be regarded as the function of trust [21,22]. Furthermore, recent studies in the transportation domain showed that policy-makers should exploit public trust resources to reinforce public demand for mitigating the risk of COVID-19 infection in public transport [23]. Notably, China is a highly populous country, and its cross-regional differences in trust are found to reflect regional differences in education, marketization of economies, urbanization, population density, and transportation facilities [24]. Meanwhile, trust to administrative institutions in China mainly comes from the satisfactory of institutional performance [25]. The levels of public trust for government can be advanced by increasing levels of financial affluence [26]. Under the social context of China, the correlation between local attachment, community trust, and civic participation in disaster-threatened areas can provide support for the construction of the community resilience system and improve the comprehensive ability of communities to resist disaster and reduce loss [27].

Furthermore, social trust depends on the individual’ s perception towards the objects (trustor) to some extent [28]. For example, in the public policy field, Siegrist and Cvetkovich et al. found that the public who perceive state and local government as being more knowledgeable about hazards, or more responsible for protecting, always have higher level of trust in the government’ stability and willingness [29]. Moreover, trust is a key to increase the acceptance by the public of the policy and the management of government and related institutions [30,31]. The public trust of the efficiency in hazard management and environmental controversies could reduce influence in rapid post-disaster recovery [32]. Trust is also an important factor to promote the energy-related initiatives [33]. Recent researches studied the distinction of trust from different cultural groups, for the reason that the farmer as internal migrants constitute the low-income group or the disadvantaged group under the Chinese special (hukou) system [8,34,35].

2.2. Community Trust and Related Factors

From a specific territorial context, community trust refers to an individual’s belief in other individuals and organizations in the community [11,36,37], and is a crucial factor in enhancing a consumer’ s intention, in stimulating developments in virtual communities (on-line) [38,39], and for ensuring effective communication during a pandemic [40]. From the aspect of the social environment, the residents in a community try to make sense of the social environment in which they live, and try to integrate the dominant groups who share similar social characteristics such as the level of education, financial status, lifestyle, and so on [41,42].

Community trust is an essential part of a community, and can be viewed as the resources embedded in the social network and which are available to the individuals in the community [43]. Community trust can facilitate social contacts like cooperation and reciprocal relationships as well as strengthen ties in social networks; in contrast, weak ties in social networks of the community may lead to disparities in the society [44]. It has been reported that community trust as a social indicator is related significantly to community engagement [36]. However, few studies have focused on the community trust of a group in the context of the effect of the built environment. Concerning the built environment, it has been found that there is a positive and substantive meaningful relationship between community trust and household health [45]. Moreover, it has been found that establishing information sharing, cooperation, and constraint mechanisms to improve community trust can have positive impacts on the public’s participation behavior within a community [27]. Some studies have indicated that residential satisfaction is associated positively with community trust [46,47].

2.3. Community Trust of Low-Income Groups

Urban poverty in China is significant and the concentration of the low-income groups is always at a neighborhood scale. The low-income group consists primarily of some specific groups, such as migrants moving from rural to urban areas, and unemployed local households [13]. With respect to the low-income group, it is a hukou system, which determines citizenship in the Chinese context. Without a local hukou certificate for the urban area, the rural migrant and urban-to-urban migrant are deprived of various social, economic, and political rights [48]. Additionally, migrants in big cities experience social isolation, limited access to benefits, and unequal working and financial conditions [49]. Wu et al. found that low-income neighborhood migrants more frequently interact with the local residents, though some local hukou residents in low-income neighborhoods are not accustomed to the different cultures and lifestyles of the migrants [50]. Moreover, the poor rural migrants who are aware of their limited resources tend to engage solely in informal social interactions. To improve cohesion, there is a need to better understand the mechanism of community trust in low-income communities [50].

Community trust between different social groups is often considered a crucial indicator of the place of belonging [51,52]. Wang et al. found greater intergroup trust between native and immigrant people in varied communities with higher immigrant population density [53]. Van Kempen et al. suggested that there was a spatial dimension at the community level for the poor with limited resources and less opportunities and government policy has promoted the concept of ‘mixed neighborhoods’ to disperse the concentration of the low-income groups [54]. Nevertheless, Wu et al. found that the neighborhood effect was insignificant in low-income groups in one specific case in China [50].

3. Data and Methods

3.1. Study Area

The study was conducted in nine cities in the Pearl River Delta (PRD) region (Guangzhou, Shenzhen, Foshan, Dongguan, Zhuhai, Jiangmen, Zhongshan, Huizhou, and Zhaoqing), which are located in Central and Southern Guangdong Province (Figure 1). The PRD region is one of the most dynamic economic city clusters in China. Under the auspices of the large bay area of Guangdong development plan issued by the State Council, the PRD joint Hong Kong and Macao special administrative region is tasked with building a world-class bay area of superior infrastructure and technological innovation. The PRD region is the center of economic activity and population agglomeration given its advantageous location and cooperative foundation [55].

Against a background of rapid economic development and growth of the population, the PRD region has experienced dramatic structural and spatial transformation [56]. Due to the differences in urban characteristics and development orientation, polarization phenomena do exist, such as gaps in the economic strength between cities and an unbalanced development within the cities of the region [57,58].

3.2. Data

The data used in this study originate from the China Labor-force Dynamics Survey (CLDS) 2016, providing a tracking database at the individual, household, and community level (available online: http://css.sysu.edu.cn/Data, accessed on 1 May 2019). CLDS, a biannual follow-up survey of village dwellings and rural areas in China, was conducted by the Center for Social Survey of Sun Yat-sen University (CSS). It established a comprehensive database of labor on demographic characteristics, socioeconomics, housing conditions, and community contexts in the survey, which has been widely applied in many studies [59]. In this study, we used the individual-level database as the original data and selected the data of the PRD from the database and defined the “personal income levels in 2015 that were lower than the personal average income for 2015 in the cities” sample as the low-income group. Finally, the present study data included 1519 valid samples, including data for 1015 local residents and 504 migrants.

Table 1. Sociodemographic characteristics of the entire sample.

All Participants		Local Residents	Migrants
Sample Size	1519	1015	504
Age/year	41.54 (std = 13.96)	44.27 (std = 13.75)	36.02 (std = 10.77)
Gender (%)
Male	48.64	49.08	47.66
Female	51.36	50.92	52.34
Education (%)
Elementary school or below	23.25	24.39	20.90
Secondary school and senior high school (secondary specialized school)	63.80	61.32	68.75
College or university	12.69	13.99	10.16
Master’s or above	0.26	0.30	0.19
Marital Status (%)
Married	77.91	79.79	74.02
Not married, divorced, or widowed	22.09	20.21	25.98
Employment status (%)
Employed (Including temporary work)	90.54	89.79	91.99
Unemployed/unemployed, laid-off/retired	9.46	10.21	8.01
Hukou Status (%)
Local hukou	66.65	-	-
Non-hukou migrants	33.35	-	-
Personal income in 2015/(Yuan/Year) (%)
Below 25,000	53.11	60.64	38.28
25000–50,000	43.07	37.12	54.69
50,000 and above	3.82	2.44	7.03

presents the sociodemographic characteristics of the subjects studied. The average age of the total sample was 41.54 years old, amongst which the local residents accounted for a relatively high proportion (66.65%), the ratio of male to female was 48.64: 51.36, the number of married persons reached 77.91%. In terms of the social and economic characteristics, 12.95% of those sampled had a bachelor’s degree (college) or above, those employed (including temporary work) accounted for 90.54%, and the total personal income ranged from zero to 25,000 yuan, accounting for 53.11%.

With respect to the socio-demographic characteristics of the two-component samples in the low-income group, the mean age of local residents (44.27) was slightly higher than that of the migrants (36.02). Regarding the socioeconomic characteristics, the employment ratio of migrants (91.99) was higher than that of local residents (89.79), and the two-component samples were mainly those with an education level of junior high school and senior high school (technical secondary school). In terms of the household incomes, the household incomes of local residents which were 60,000 to 250,000 yuan accounted for 64%, while the majority of household incomes for the migrants ranged from 25,000 to 50,000 yuan, accounting for 54.69%.

3.3. Variables

In this study, 1519 valid samples for the “low-income group” were obtained, including 1015 local residents and 504 migrants. The data processing and analysis were conducted using SPSS 25.0 and STATA 16.0 statistical software. The selection of the variables and indices for the present study were performed as follows.

3.3.1. Dependent Variable

The dependent variable was the community trust for the low-income group. Referring to the research conducted by Wu et al. [60], the measurement of the community trust for the low-income group was based on using “the level of trust to the community, the neighborhood and other residents”. The 5-point Likert scale was used in the study whereby higher points in the score indicated a higher level of community trust. For example, 1 point equates to “very distrustful” and 5 points means “very trustful”.

3.3.2. Independent Variables

The study selected social environmental factors and urban environmental factors as independent variables to explore their influence on the community trust among the low-income group. Three indicators were used for measurement with respect to the social environmental elements, namely, social contact, community participation, and sense of security. First, to assess social contact we used the number of friends or acquaintances who can give support and help locally. Second, for community participation we adopted the frequency of participation in nine group activities (neighborhood committees, social work organizations, ownership committees, leisure/recreation/sports clubs/salon organizations, learning/training institutions, township associations, clan organizations, public welfare/social organizations/volunteer groups, and religious organizations). The 5-point Likert scale was used to count the frequency of participation and the higher scores represented higher levels of community engagement for each item with a total score of 9–45 points. The possibility of encountering six problem issues (unemployment, crime, terrorist attacks, access to fake or shoddy food, infection with an infectious disease, and environmental pollution) was applied to measure the sense of security based on using the Likert scale where higher scores indicated higher levels of a sense of security with values ranging from 1 to 5 points with a total score of 6–30 points.

In addition, for the urban spatial elements, we included the urbanization rate, the per capita green space area, the bus line network density, the POI (point of interest) density, the intensity of land development, and the population density. The data of urbanization rate, urban green space area, the length of public transportation operation route, urban built-up area, administrative region area and population density were all originated from the Guangdong Statistical Yearbook in 2016. Furthermore, the data of the urban POI quantity was extracted by reference to the map in 2015 from the river map (http://www.rivermap.cn/index.html, accessed on 1 May 2019), which reflected the quality of the residents’ life, mainly considering the types such as entertainment, catering, education, medical, catering, tourist attraction, and so on. The number of POI per 10,000 people was calculated within the 1000 m shape element buffer zone of each community with the POI data and the number of community’s permanent population from neighborhood committees. Additionally, the population density is based on the communities’ administrative area. The specific indicators are listed in

Table 2. Urban space dimension index.

Index	Definition
Urbanization rate	The proportion of the urban population in the total population
Per capita green space	The per capita content of the urban green space area
Bus line network density	The ratio of the length of the public transport operation routes to the urban built-up area
POI density	The ratio of the number of urban POI to the urban administrative area
Land development intensity	The proportion of the urban built-up area to the administrative area
Population density	Number of people living on land per square kilometer area of land

.

3.3.3. Control Variables

The control variables in this study are indicators of the sociodemographic characteristics, and include nine aspects: gender, education, marital status, political status, employment, hukou status, age, level of English ability, and health status. Among them, the employment status means the latest type of a job divided into five categories, that is, public ownership type work, private enterprise or organization, individual industrial and commercial households, farming, and freelance workers. Further, the health status includes self-rated health and hospitalization. For the ranking of self-rated health, 1 point corresponds to very unhealthy and 5 means very healthy; hospitalization refers to hospitalization diagnosed by doctors in the past two weeks.

3.4. Model Specification

The community trust of the low-income group is influenced by the sociodemographic characteristics, the social environmental dimension, and the urban space dimension [42,43,44,45,46,47,48,49]. Before the analysis, the influence of different variables of community trust in the low-income group was first discussed by the null model, that is, constructing a null model of community trust containing only the dependent variable and calculating the intra-class correlation (ICC) [61]. The ICC was calculated as follows:

$$\mathrm{ICC}=\frac{{\mathsf{\sigma }}_{\mathrm{b}}^2}{{\mathsf{\sigma }}_{\mathrm{w}}^2+{\mathsf{\sigma }}_{\mathrm{b}}^2}$$

(1)

where ${\mathsf{\sigma }}_{\mathrm{b}}^2$ represents the variance between cities and ${\mathsf{\sigma }}_{\mathrm{w}}^2$ represents the individual variance within cities. By measuring the ICC values of the total sample, ICC = 0.0376 means that using the multi-layer linear regression model is not significant, so it is more appropriate to use the linear regression model to measure the mechanism of influence of community trust for the low-income group [62]. The expression is as follows:

$${\mathrm{Y}}_{\mathrm{ij}}={\mathsf{\alpha }}_1+{\mathsf{\eta }\mathrm{X}}_{\mathrm{j}}+{\mathsf{\beta }}_1{\mathrm{Z}}_{\mathrm{ij}}+{\mathsf{\gamma }}_1{\mathrm{W}}_{\mathrm{ij}}+{\mathsf{\mu }}_{\mathrm{ij}}+{\mathsf{\epsilon }}_{1\mathrm{ij}}$$

(2)

where ${\mathrm{Y}}_{\mathrm{ij}}$ represents the level of community trust of a low-income individual i of city j; ${\mathrm{X}}_{\mathrm{j}}$ represents the urban space factor variable of city j; ${\mathrm{Z}}_{\mathrm{ij}}$ represents the variables for the sociodemographic characteristics of resident i of city j; ${\mathrm{W}}_{\mathrm{ij}}$ represents the variable for the social environment of resident i of city j; ${\mathsf{\alpha }}_1$ represents the intercept; $\mathsf{\eta }$ is the total effect of the independent variables; ${\mathsf{\beta }}_1$ represents the coefficient of the sociodemographic characteristics and ${\mathsf{\gamma }}_1$ represents the coefficient of the social environment; ${\mathsf{\mu }}_{\mathrm{ij}}$ represents the residual of the sociodemographic characteristics; and ${\mathsf{\epsilon }}_{1\mathrm{ij}}$ represents the residual of the social environment.

4. Results

4.1. Effects of Urban Environment on Community Trust

The results of the regression analyses indicate that both the urbanization rate (−0.006, p < 0.05) and the intensity of land development (−0.013, p < 0.05) are negatively associated with the community trust of the low-income group. Specifically, a one-point increase in the rate of urbanization produces a 0.006-point decrease in community trust. Additionally, for an increase of one point in the intensity of land development, the community trust decreased by 0.05%. This demonstrates that urban development hinders the progression of community trust with the residents feeling stranger and stranger with each other due to the enhanced population flows and increasing growth in the population [63]. Furthermore, the population density has a significant negative impact on community trust, on account of the fact that it reduces the level of the community trust of the low-income group. An increase in the population density will also impact negatively on infrastructure and public service facilities per capita to a certain extent. Moreover, there would be an increasing possibility of friction occurring among the community residents regarding their “ownership” of the public resources [46] (

Table 3. Regression results for social trust in the entire sample.

	Coeff.	S.E.	t-Value	p-Value
Sociodemographic characteristics
Gender	0.013	0.043	0.31	0.760
Secondary/senior high school	0.034	0.057	0.60	0.551
College or university	0.129	0.094	1.36	0.173
Master’s or above	−0.929	0.415	−2.24	0.025 **
Married	0.036	0.059	0.61	0.539
Party member	0.134	0.109	1.23	0.221
Employment status	−0.040	0.071	−0.56	0.577
Private enterprises or organizations	−0.130	0.074	−1.76	0.079 *
Individual industry	−0.111	0.088	−1.26	0.207
Farming	0.034	0.095	0.35	0.726
Freelance workers	−0.037	0.094	−0.40	0.692
Agricultural hukou	−0.021	0.057	−0.37	0.713
Age	0.009	0.002	4.05	0.000 ***
English level	−0.006	0.069	−0.09	0.929
Self-rated health	0.110	0.026	4.22	0.000 ***
Hospitalized for the last two weeks	0.039	0.080	0.48	0.629
Social Environment Dimension
Social contact	0.003	0.001	2.89	0.004 ***
Community participation	0.015	0.015	0.99	0.323
Sense of security	0.007	0.005	1.52	0.130
Urban Space Dimension
Per capita green space	−0.009	0.022	−0.41	0.685
Population density	−0.000	0.000	−2.32	0.020 **
Urbanization rate	−0.006	0.003	−2.36	0.018 **
Bus line network density	0.001	0.002	0.26	0.794
Land development intensity	−0.013	0.005	−2.45	0.015 **
POI density	0.005	0.002	2.62	0.009 ***
Constant	3.064	0.495	6.19	0.000
Sample size	1519

Note: * p < 0.10; ** p < 0.05; *** p < 0.01.

).

The POI density (regression coefficient is 0.005, p < 0.01) is significantly positively correlated with community trust. For an increase of one point in the POI density, the community trust increased by 0.005. The POI density indicates an increase in the community service facilities, which effectively plays an important role in improving the level of community trust of the low-income group [5,64,65]. When considering the factors associated with the social environment dimension, the results indicate that social contact has a significant impact on community trust (regression coefficient is 0.003, p < 0.05). For an increase of one point in social contact, the community trust increased by 0.003. The more people in the low-income group that are receiving support indicates that a larger social support network has been constructed for the community, which can effectively improve the level of community trust [66,67].

When considering the sociodemographic characteristics, it can be seen that age and health status has a significant positive effect on the community trust of the low-income group (regression coefficient was 0.009, 0.110, p < 0.01). Compared with the young, the elderly are closely associated with other residents for the reason that they tend to have more leisure time and are more able to participate in community affairs and activities, which improve the level of community trust. These findings are consistent with published research findings. However, there were two indicators that were negatively associated with the community trust for the low-income group. That is, those persons with a master’s degree or above (regression coefficient was −0.929, p < 0.05) and those who work in private enterprises or organizations (regression coefficient was 0.627, p < 0.05); these two groups tend to have work pressure and are too busy to spend much time engaged in community activities, hence there is a lower level of community trust [47,67].

4.2. Local Residents vs. Migrants

For local residents (

Table 4. Regression results for community trust in local residents and migrants.

	Local Residents				Migrants
	Coeff.	S.E.	t-Value	p-Value	Coeff.	S.E.	t-Value	p-Value
Sociodemographic characteristics
Gender	0.038	0.050	0.77	0.441	0.017	0.080	0.21	0.834
Secondary/senior high school	0.006	0.067	0.09	0.927	0.015	0.105	0.14	0.886
College/university	0.092	0.112	0.83	0.408	0.116	0.174	0.67	0.503
Master’s or above	−0.542	0.460	−1.18	0.239	−2.254	0.903	−2.50	0.013 **
Married	0.016	0.071	0.22	0.826	0.081	0.112	0.72	0.470
Party member	0.228	0.120	1.90	0.058 *	−0.126	0.254	−0.50	0.620
Employment status	0.001	0.082	0.01	0.994	−0.066	0.142	−0.47	0.642
Private enterprise or organizations	0.010	0.082	0.12	0.901	−0.349	0.185	−1.89	0.059 *
Individual industry	0.053	0.106	0.50	0.616	−0.319	0.196	−1.63	0.105
Farming	0.032	0.101	0.32	0.750	0.627	0.313	2.00	0.046 **
Freelance worker	0.027	0.104	0.26	0.796	−0.157	0.220	−0.71	0.475
Agricultural hukou	0.139	0.071	1.95	0.052 *	0.019	0.133	0.14	0.886
Age	0.009	0.003	3.48	0.001 ***	0.004	0.005	0.90	0.367
Foreign language level	0.002	0.084	0.02	0.985	0.119	0.124	−0.96	0.337
Self-rated health	0.127	0.031	4.15	0.000 ***	0.071	0.049	1.44	0.150
Hospitalized in last two weeks	0.096	0.089	1.08	0.279	−0.145	0.168	−0.86	0.389
Social environment Dimension
Social contact	0.002	0.001	2.31	0.021 **	0.003	0.003	1.06	0.292
Community participation	0.003	0.019	0.14	0.892	0.037	0.026	1.43	0.152
Sense of security	0.006	0.006	1.05	0.296	0.011	0.009	1.28	0.202
Urban space Dimension
Per capita green space	0.008	0.028	0.29	0.774	0.037	0.117	0.32	0.752
Population density	0.000	0.000	0.09	0.932	−0.001	0.001	−1.21	0.228
Urbanization rate	0.007	0.003	1.97	0.050 *	−0.013	0.015	−0.83	0.407
Bus line network density	0.001	0.003	0.20	0.842	0.002	0.009	0.18	0.855
Land development intensity	0.007	0.007	0.91	0.364	−0.021	0.022	−0.96	0.337
POI density	0.001	0.003	0.44	0.661	0.008	0.008	1.09	0.278
Constant	2.847	0.712	4.00	0.000	2.774	1.587	1.75	0.081
Sample size		1015					504

Note: * p < 0.10; ** p < 0.05; *** p < 0.01.

), the regression model results indicated that in the social environment dimension, the factors associated with social contact (regression coefficient of 0.002, p < 0.05) had a significant positive impact on community trust among the low-income group. Social contact helps the residents to make more friends and establish close relationships and social circles in the community. As a result, social contact promotes community trust in the low-income group [47]. As for the urban space dimension, the urbanization rate has a significant negative correlation (regression coefficient of −0.007, p < 0.10) with the community trust of the low-income group, that is, the community trust of the low-income group decreases the higher the urbanization rate. For rapid urbanization, population mobility in the community is more frequent, and the heterogeneity of the community population increases, both of which have a negative impact on neighborhood relationships. Therefore, the community trust of the low-income group is reduced in these settings [68,69].

Among the sociodemographic characteristics, the indicators of party membership (0.228 regression, p < 0.10), agricultural hukou (0.139, p < 0.10), age (0.009, p < 0.01), and self-rated health (p < 0.01) are all significantly and positively associated with community trust among the low-income group. Among them, the low-income group with party membership is more active in community activities and social contact compared to the masses or other political factions; the low-income group with agricultural hukou has a social support network, so this group is more active in social interactions. The residents with high self-rated health scores not only have a more positive attitude towards life but also are more eager to participate in social activities, so this is beneficial to improving their level of community trust [70,71].

For the migrants (Table 4), the results of the regression model indicate that both the social environment and the urban space have no significant impact on their levels of community trust, and only the social and demographic attributes have a significant impact on community trust. Among them, those with a master’s degree or above (regression coefficient: −2.254, p < 0.05) had a significant negative correlation with private enterprises or organizations (regression coefficient: −0.349, p < 0.10) and with community trust in the low-income group, while farming (regression coefficient: 0.627, p < 0.05) was positively associated with community trust in the low-income group.

5. Discussion and Conclusions

5.1. Discussion

To achieve the goal of future cities and meet the need of livable community and faced with threats such as air pollution, water pollution, waste management, non-renewable resources consumption and so on, the government needs to place emphasis on the efficient using of urban space, enforcement of environmental protection, minimizing the consumption of essential natural resources, and mobilizing citizens’ participation on environmental projects [50]. To better enhance the quality of urban life, interventions could be made to support an equitable and accessible built environment, high quality urban life, integrated transportation, and land use [72,73,74].

Therefore, it is necessary to pay attention to the optimization of material space and social space, and to promote the formation of low-income group-friendly livable communities in community planning and governance. Specifically, in terms of material space planning, the environment should be optimized with sufficient public space, to provide the chance to communicate and establish a social network for the low-income group. Moreover, it is crucial to improve support facilities and the education and medical infrastructure so as to improve living conditions for the low-income group.

With respect to governance of the social space, it is essential to consider the maintenance of the existing networks in the community and the construction of the new community networks [75,76]. Therefore, establishing communication platforms between local residents and migrants would be helpful for improving trust amongst the residents. Moreover, the government should pay more attention to the low-income groups in urban communities and provide employment training to promote stability in employment and in income. The community as an institution should implement the government policies, provide support services, improve the social resources, and improve the quality of life for the low-income group in the community, thus enhancing the self-identification of the low-income group.

Furthermore, the planning at the community level should consider the different needs of different groups, especially for the disadvantaged. For local residents, in order to enhance their community trust, it is important to carry out community activities to maintain and expand their networks with the local residents and migrants. For the migrants, given their mobility and weak social identity, it is essential to organize community activities to promote social contact. The government should prioritize helping the migrants and instigate measures for their community adaptation and emotional integration at the community level, including the provision of better housing conditions and more accesses to support facilities. Improving the level of community trust of low-income groups is beneficial to accelerate the development of community integration and sustain livable communities in the future [13].

5.2. Conclusions

With the rapid development of urban agglomeration in the PRD, there has been a clear and radical transformation and reconstruction of urban communities since the reform and opening up of China [60,77]. In addition, the problem of community integration of low-income groups in cities has also attracted much attention. Research on community trust of the low-income groups is of great significance for the protection of the interests of the disadvantaged and for the building of livable communities [78]. In this context, the present study focuses on the low-income group in the PRD region, examining the influence and the mechanism of the community trust from the social environment dimension, the urban space dimension, and the sociodemographic characteristics dimension. In addition, this study explores and discusses the differences in community trust between the local residents and migrants. The study reveals that the mechanism of community trust of low-income group has the following associations:

(1)

Regarding the social environment dimension, social contact can directly enhance the community trust of the low-income group. POI density is positively associated with community trust, while the population density, the urbanization rate, and the intensity of land development are negatively associated with community trust. Moreover, age, self-rated health, and local residency exert significant positive effects on community trust, while indicators such as those with a master’s degree or above, who depend on private enterprise or an organization for employment, are significantly negatively associated with community trust.

(2)

For local residents, social contact and the indicators of social and demographic attributes such as political party membership, agricultural hukou, age, and self-rated health are all significantly positively associated with community trust. However, the urbanization rate, which is an urban space factor, has a significant negative impact on community trust. It was found that local residents have a high level of community trust on account of this group having accumulated over the long-term social contact networks. However, rapid urbanization brings about an increase in the heterogeneity of community populations, which reduces the level of community trust of local residents to some extent.

(3)

For the migrants, from sociodemographic characteristics dimension, working in private enterprises or organizations and in agriculture (farming) has a significant impact on community trust. This is because migrants in the low-income group showed a low level of sense of social security and were lacking a social network relationship. Therefore, it is important to construct strong social networks for migrants. Hence, a relatively stable economic income will enhance their emotional connections to their communities now and in the future.