Sustainable Community Transformation and Community Integration of Agricultural Transfer Population—A Case Study from China

Abstract

Urban-rural integrated communities (URICs) are transitional areas for agricultural transfer population (ATP) in the process of urbanization in China. In the current urban renewal context, the demolition and renovation of communities often result in ATP living in a precarious situation and being marginalized in the city. Sustainable urban renewal should change this situation, take the transformation of URICs as a breakthrough, and promote the urban integration of the ATP. Based on the survey of the National Health Commission of China in 2017, this paper investigates the effects of community participation and community identity on community integration, using the ordered probit model with data of the ATP living in URICs. The results show that both community identity and community participation positively influence community integration, and there are intergenerational differences. The mediating effect test shows that community identity plays a mediating role in the process of community participation, influencing community integration. The findings of the study provide possible ideas for the practical promotion of community integration and urban integration of the ATP, facilitate the implementation of sustainable urban regeneration to reduce the costs of citizenship, and maximize the benefits of the cumulative effects of urbanization for all segments of the population.

1. Introduction

As urban-rural integration becomes a national development strategy, China enters a strategic opportunity period for urbanization, and creates conditions for agricultural population migration and mobility.

The movement of agricultural populations from villages to cities is a common phenomenon in the process of urbanisation around the world. In early 20th century in America, the mechanization of agricultural production, land annexation and high farmland rental costs led to the exodus of thousands of farmers to manufacturing cities in the Northeast and Midwest. As a result, urbanised migration contributed to the growth of cities such as Detroit, Cleveland and Chicago [1]. Rural populations in Southeast Asia may hope to improve their socio-economic status by living in cities. However, upon entering the city, they may have to work in jobs they do not like, or even suffer from overwork and abuse due to their irregular status [2]. For political-economic reasons, rural migrants from Latin America who move to the cities face inadequate sanitation and shelter, and even a lack of adequate water and food [3]. China’s agricultural transfer population (ATP) is a special group created in the context of urban-rural dualistic governance. At the beginning of China’s reform and opening up, there was an imbalance between urban and rural economic development [4]. With the adjustment of agricultural policies, surplus laborers who were originally engaged in agricultural activities left their hometowns and moved to the cities in search of jobs. These people moving into the city were engaged in various occupations such as waiters, production workers, technical workers, construction workers, general clerks, individual owners, sanitation workers, etc., forming an ATP group. As they move to cities, their accommodation needs to be urgently addressed [5].

The urban-rural integrated communities (URICs) become an urban residential option for the ATP. URIC was formed as a result of urban sprawl, where the land around the edge of the city was heavily expropriated and the original rural settlements were surrounded. These original residents were gradually incorporated into the urban areas. However, except for some who were transformed into urban residents by expropriation, the rest still belong to farmers. The nature of these lands changed from agricultural land to mainly residential and collective land, and the URICs where the ATP lives together with residents are formed [6]. Most URICs are located near commercial areas, close to workplaces, with easy access to public facilities and good services. The above-mentioned advantages have led to the formation of a family economy. The original residents provide cheap rental accommodation to migrant workers, and the URICs have gradually evolved into a major settlement for ATP [7].

In recent years, to promote the economical and intensive use of land resources, the government has concentrated on the demolition and renovation of URICs. [8]. This has resulted in a large number of ATP living in URICs having to move from one community to another, making it difficult for them to truly integrate into urban life and keeping them marginalized [9]. The lack of a stable living situation and the dichotomy between urban and rural areas make them not fully recognize the community as a place to call home. Also, some URICs have fallen into a state of disorder to varying degrees, due to the lack of effective rehabilitation and reliable policy protection [10]. The construction renovation of the community without including the migrant population as residents in planning will not only increase the cost of community transformation, but also leave behind hidden dangers of urban renewal. In the new stage of renewal for URIC, the promotion of benefit-sharing, and the community integration of the ATP determine the effectiveness of the transformation [11].

The community integration of the ATP is an important goal of the Chinese government’s “people-oriented” urbanization. If the transformation of URICs cannot take up the task of ATP’s citizenship, it is destined to be unsustainable, the citizenship strategy will not be completed, and China’s new type of urbanization will become a castle in the air. In such a context, this paper uses data from China Migrants Dynamic Survey (CMDS) in 2017 to analyse the factors affecting the community integration of ATP and their intergenerational differences, and the mediating effects to test the pathways of the integration process. The study aims to use the transformation of URICs as a breakthrough to promote the integration of the ATP into the city, thereby truly advancing citizenship. The study can provide empirical evidence for government departments to formulate policies and provide a reference for decision-making to promote the citizenship process.

2. Literature Review

2.1. Community Integration of the ATP

Although the community integration of ATP is a typical Chinese problem, the Western migration theory is still informative. Western explanations of migrant integration are mainly from the perspectives of human capital characteristics [12], social networks [13] and institutional policies [14], revolving around migrants’ education level, work experience, social connections and employment security. As a local issue, Chinese scholars focus more on the process and outcome of community integration in the context of the national reality. The community, as a transitional field for the citizenship of ATP, provides a set of orderly norms for community members to follow. As a disadvantaged group restricted in many ways of household registration and status, the ATP passively reshapes their own values and lifestyles in the community arena [15]. Through active participation in community activities, access to community services and concerning decision-making procedures, ATP gradually strengthens the community identity [16]. When communities become well-managed groups and have the ability to do practical things for the residents, they can give the migrant population a sense of belonging, thus helping them to achieve the re-socialization process of integrating into the community and then into the city [17].

2.2. Community Identity of the ATP

From the aforementioned analysis, it can be seen that the process of community integration of the ATP is closely related to the community identity, and the identity exists in the spatially close circle of acquaintances [18]. However, there are some contradictions between the community membership and the reality of the ATP, which come from society, others and within the ATP group [19]. On the one hand, ATP has a low self-identification. They live with the original residents of the community, but the psychological alienation based on geographical ties, the boundary of interpersonal interactions, and doubts about their citizenship make it difficult to build their self-identity [20]. On the other hand, local residents do not fully identify with the membership of the ATP. The urban-rural divide is often associated with the personal quality of the population, with a bias that agricultural workers are generally insufficiently educated and therefore unable to fully understand policy rules [21]. The accompanying poor occupational prestige, management difficulties and security problems make it even harder for locals to form a community identity for the migrant population [22]. Moreover, some ATP fail to obtain urban residence and are therefore excluded from social rights such as housing, social security and community benefits [23].

2.3. Community Participation of the ATP

Inadequate community identity leads to undesirable consequences When faced with community public affairs, ATP often makes selfish choices and finds it difficult to form effective cooperation with residents and the government. Some scholars have pointed out that the ineffective collective decision-making and lack of transparent planning processes have led to potential social conflicts and inequalities [24]. Also, successful governance experiences have shown that community participation can generate positive feedback from the migrant population [25]. For example, in Beijing’s Zhonggu Lou neighborhood renovation project, residents participated in the conversions, expressed their demands and fought for their rights through new community networks such as microblogs and WebGIS [26]. In Beijing Pi Village, a community participation platform called “Worker’s Home” was set up to enable migrant workers to defend their rights, and share the fruits of community development [27]. For migrant working families, especially those with children or living with elderly parents, participation in activities can provide community support, thus compensating for the lack of policy-based services due to household registration restrictions [28]. Although the impact of informal community participation like this is limited, these public activities do benefit the ATP.

In summary, community integration and its closely related issues of community participation and community identity are both hot research topics for ATP. Moreover, the existing studies have selected multiple factors from various aspects such as housing, household registration and social security. In contrast to previous literature, the possible innovations of this paper are as follows. First, the research perspective is novel. It is based on URICs in the process of urban renewal. In the analysis, the mediating role of community participation in influencing community integration is taken as the central focus, while intergenerational differences are taken into account. Second, the empirical strategy is rigorous. Data was collected from authoritative and reliable surveys, with wide and representative sample coverage. The model is also tested through replacement variables and replacement regression methods to ensure the robustness of the findings. Third, the argumentation process is complete. The mechanism of the role of community integration is explored based on the mediating effects model.

3. Theoretical Analysis

There are many schools of thought to explain the community integration of migrant populations, and the most common ones are social identity theory and push-pull theory, in addition to social support theory [29] and community field theory [30].

3.1. Social Support Theory

Social support refers to the help provided by people who play a key role in an individual’s life. Individuals develop a variety of social relationships through their participation in activities with other individuals, groups, and communities [31]. The community, as the place where migrants settle, is also the place where migrants receive social support. While the ATP and the urban residents in the same community lack substantive interaction, it is difficult for ATP to gain recognition from other residents and establish new social support networks, which become the migrants’ barriers to their integration into the community [32]. Conversely, if the migrants build a support network with the group, it facilitates their mutual identification with group members and enhances their integration [33]. The migrants can use participation in community activities as a platform, thereby increasing community acceptance and achieving community integration through the acquisition of horizontal social capital and the extension of social networks, and finally complete overall urban integration [34].

3.2. Push-Pull Theory

The push-pull theory is widely used to analyse the phenomenon of urban-rural migration [35]. The pull of ATP to integrate into urban communities comes from stable accommodation, easy access to public services and good living conditions. But at the same time, unfamiliar environments, fierce competition, and cultural and customary differences push them away [36]. Push and pull factors vary between generations [37]. Most of the older migrant agricultural workers cling to traditional thinking. After a hard struggle in the city, they generally choose to return to their hometowns and build houses with the income they have earned outside. The new generation of ATP, on the other hand, have grown up in the city, are generally better educated and more adaptable and receptive than their parents, and thus have a stronger inclination towards the city and a greater desire to integrate [38].

3.3. Social Identity Theory

Social identity theory discusses the relationship between individuals and groups. The theory was developed in the 1980s by Tajfel et al. [39], and has gradually become an important theoretical tool for exploring the problems of socially mobile people and disadvantaged groups [40]. The theory argues that social networks formed through collective participation can increase an individual’s access to information and thus to more social support resources than social activities between individuals. The increase in individual self-worth in turn has a positive impact on collective identity [41]. The application of this theory can explain the phenomenon that immigrant community involvement influences place attachment and community identity, and that community participation is one of the intermediary mechanisms that influence place attachment [42]. Similarly, social status perceptions influence urban identity, and there is a potential mediating effect between the two [43].

Based on the above analysis, we put forward the following hypotheses, and the corresponding relationships are shown in Figure 1.

Hypothesis 1 (H1).

Community integration of the ATP is influenced by factors such as community identity and community participation, and there are intergenerational differences.

Hypothesis 2 (H2).

Community identity of the ATP is a mediating variable of community participation affecting community integration.

4. Methodology

4.1. Data Source

The data comes from the 2017 CMDS. CMDS is a nationwide survey conducted by the National Health Commission of China to understand the survival and development of the mobile population, and the trends and characteristics of migration. The trained interviewers use the smartphone or PAD to carry out the face-to-face surveys. The questionnaires are divided into 4 types: Modules A-D. Among them, Modules A and C focus on the transfer population from the individual level. And Module C focuses on 8 representative cities on the basis of Module A, and conducts a more detailed investigation of the transfer population. Also considering the availability of data, we select the data of Module C as the research sample. The questionnaires use the Probability Proportionate to Size Sampling method in 8 target cities (districts), Jiulongpo District in Chongqing, Guangzhou, Qingdao, Zhengzhou, Suzhou, Changsha, Xishuangbanna State and Urumqi. People being investigated are the migrant population who have lived in the survey place for more than 1 month, with non-local household registration, and aged 15 years and above in May 2017. Finally, we obtain a total of 13,998 samples of the questionnaire.

The samples are screened according to the condition of “current housing status”. The “purchased commercial houses” and “self-built houses” are classified as “self-owned houses”; “low rent houses”, “public rental houses” and “other indemnification houses” are classified as “policy houses”; “renting private houses”, “borrowing houses”, “places of employment” and “other informal places” are classified as “renting houses”. The last group of renting houses is defined as ATP living in URICs, and is the respondent of this study. After the preliminary screening of the questionnaire, the number of effective samples is 10,015.

4.2. Variable Selection

4.2.1. Dependent Variable: Community Integration

In terms of measuring the community integration of ATP, most studies have used questionnaires, in-depth interviews and field observations to investigate psychological integration, sense of belonging and psychological distance. Regarding existing studies [44] and the question “I would like to integrate and become a member of the local population” provided in the CMDS of 2017, the responses are classified as “totally disagree”, “disagree”, “basically agree” and “totally agree”, which are assigned values 1 to 4 accordingly as the measurement of willingness to integrate into the community. The measure of “community integration” is a measure of the whole ATP’s willingness to integrate into the community in this study.

4.2.2. Independent Variables: Community Participation and Community Identity

(1)

Community identity

Referring to the relevant findings of existing studies on the formation of community identity and its role in the group [45], this study selects 4 explanatory variables of self-identity, local acceptance, organizational identity and social identity to examine the community identity of the ATP from multiple perspectives.

(2)

Community participation

Referring to existing empirical studies on the set of explanatory variables related to residents’ community participation behaviour, and taking into account the study purpose of [26], we select 4 indicators, namely making suggestions, participating in discussions, participating in activities and community concern, to measure the community participation of the ATP.

4.2.3. Other Control Variables

Drawing on the content of the CMDS and combining it with the existing literature [46], this paper controls for variables that may affect the community integration of ATP, including personal characteristics such as gender, age, education level, number of household members and marriage status. The statistical descriptions of the variables and their subdivisions are shown in

Table 1. Description of variables and indicators.

Variables	Indicators	Indicator Meanings	Ave	S.D.	Min	Max
Community integration	/	I would like to integrate with the local people in the community. Totally disagree = 1, disagree = 2, basically agree = 3, completely agree = 4	3.27	0.625	1	4
Personal characteristics	Gender	Male = 0, female = 1	0.48	0.500	0	1
	Age	Actual age	39.51	9.966	20	86
	Education level	Below junior high school = 1, high school/secondary school = 2, junior college = 3, university undergraduate and above = 4	The value = 1 (58.31%) The value = 2 (26.22%) The value = 3 (10.39%) The value = 4 (5.08%)		1	4
	Number of household members	Number of household members	3.01	1.216	1	10
	Marriage status	Unmarried/cohabiting = 1, divorced/widowed = 2, married = 3	The value = 1 (21.66%) The value = 2 (76.25%) The value = 3 (2.09%)		1	3
Community identity	Self-identity,	I feel like a local already. Totally disagree = 1, disagree = 2, basically agree = 3, totally agree = 4.	2.79	0.751	1	4
	Local acceptance	I think the locals would like to take me on board. Totally disagree = 1, disagree = 2, basically agree = 3, totally agree = 4.	3.19	0.619	1	4
	Organizational identity	Interact with people in my spare time. Very little interaction with people = 1, with fellow countrymen = 2, other migrants = 3, other locals = 4.	The value = 1 (3.77%) The value = 2 (29.95%) The value = 3 (50.26%) The value = 4 (16.02%)		1	4
	Social identity	Apply for a personal social security card. Did not do it, never heard of it = 1, did not do it but heard of it = 2, have done it = 3.	2.30	0.742	1	3
Community participation	Making suggestions	Give advice or monitor management in the community. No = 1, occasionally = 2, sometimes = 3, often = 4.	2.95	0.270	1	4
	Participating in discussions	Comment on national affairs and social events online and participate in discussions. No = 1, occasionally = 2, sometimes = 3, often = 4.	2.93	0.298	1	4
	Participating in activities	Active participation in volunteer activities. No = 1, occasionally = 2, sometimes = 3, often = 4.	2.40	0.533	1	4
	Community concern	Concern about the changes in the community. Totally disagree = 1, disagree = 2, basically agree = 3, totally agree = 4.	3.29	0.612	1	4

Note: (1): The number of observations is 10,015. (2): In the columns of descriptive statistics, the indicators of educational level, marital status and organizational identity are reported with the use of frequencies.

.

4.3. Model Construction

4.3.1. Ordered Probit Model

Given that the values of community integration indicators are restricted ordinal data with discrete values, it is not possible to use ordinary linear regression models for analysis, so a multivariate ranking selection model is required. In contrast, ordered probit model (Oprobit) is a typical regression with a clear ordinal relationship between the terms. The core idea of which is to study the changing pattern of unobservable latent variables by modelling the observable ordered data [47]. Therefore, the potential willingness to integrate into the community can be considered as a latent variable $y^{*}$ and studied using the Oprobit model. The linear equation of $y^{*}$ can be expressed as:

$$y_i^{*}={\beta }_1{x}_{i1}+{\beta }_2{x}_{i2}+\eta x_i^{\prime }+{\epsilon }_i,i=1,2,...,N$$

(1)

where $y_i^{*}$ is an indicator of community integration, $x_{i1}$ represents community identity, $x_{i2}$ represents community participation, $x_i^{\prime }$ is a series of variables affecting community integration, ${\beta }_1$, ${\beta }_2$ and $\eta$ are the coefficients of the variables, and $\epsilon$ is the residual term. Although $y_i^{*}$ is an unobservable latent variable, it is related to $y_i^{}$, a series of observable ordered series as follows:

$$y_i=F(y_i^{*})=\{\begin{array}{l}1,y_i^{*}<{\alpha }_1\hfill \\ 2,{\alpha }_1\le y_i^{*}<{\alpha }_2,\hfill \\ ...\hfill \\ J,{\alpha }_{J-1}\le y_i^{*}\hfill \end{array}$$

(2)

where $y_i^{}$ is a discrete array $\{1,2,...,J\}$, representing the community integration intentions of the ith migrant; $a_1<a_2<...<a_J$ is the cut-point parameter to be estimated, and $y$ is divided into $J$ intervals. Thus, the probability that the ith observation $j$ of $y$ falls into an interval can be expressed as:

$$P(y_i=j)=\{\begin{array}{l}F({\alpha }_1-{\beta }_1{x}_{i1}-{\beta }_2{x}_{i2}-\eta x_i^{\prime }),j=1\hfill \\ F({\alpha }_j-{\beta }_1{x}_{i1}-{\beta }_2{x}_{i2}-\eta x_i^{\prime })-F({\alpha }_{j-1}-{\beta }_1{x}_{i1}-{\beta }_2{x}_{i2}-\eta x_i^{\prime }),2\le j\le J-1\hfill \\ ...\hfill \\ 1-F({\alpha }_{j-1}-{\beta }_1{x}_{i1}-{\beta }_2{x}_{i2}-\eta x_i^{\prime }),j=J\hfill \end{array}$$

(3)

Then, the dependent variable is replaced with $y_i^{}$ to construct the Oprobit model:

$$y_i=F({\beta }_1{x}_{i1}+{\beta }_2{x}_{i2}+\eta x_i^{\prime }+{\epsilon }_i),i=1,2,...,N$$

(4)

According to Equation (4), the log-likelihood function corresponding to the ith observation $j$ is:

$$\ln L={\displaystyle \sum _{i=1}^N{\displaystyle \sum _{j=1}^J\ln [}}F({\alpha }_j-{\beta }_1{x}_{i1}-{\beta }_2{x}_{i2}-\eta x_i^{\prime })-F({\alpha }_{j-1}-{\beta }_1{x}_{i1}-{\beta }_2{x}_{i2}-\eta x_i^{\prime })]$$

(5)

The coefficients (${\beta }_1$, ${\beta }_2$, $\eta$, ${\epsilon }_i$) of the Oprobit model can be obtained through the maximum likelihood estimation, which is the unbiased, consistent estimate of the corresponding parameter in Equation (1).

4.3.2. Test for Mediating Effects

Based on the proposed framework, a linear probability model (LPM) is used to test the mediating role of community identity between community participation and community integration [28]. The basic form of the linear probability model is:

$$y_{\mathrm{i}}={\eta }_1{x}_{\mathrm{i}1}+{\eta }_2{x}_{\mathrm{i}2}+\mu$$

(6)

where $y_i$ is the dependent variable, $x_{i1}$ and $x_{i2}$ are the independent variables, ${\eta }_1$ and ${\eta }_2$ are the corresponding coefficients of the independent variables, and $\mu$ is the random effect of the unobserved factor.

The dependent variable (community integration) is first regressed onto the independent variable (community participation) using the LPM. Next, the mediating variable (community identity) is regressed onto the independent variable (community participation) using the LPM. Finally, the LPM s used to regress the dependent variable (community integration) on both the independent variable (community participation) and mediating variable (community identity). The mediating effect occurs when the indicator of community integration is influenced by both the indicators of community participation and community identity, as well as when the mediating variable (community identity) is influenced by the independent variable (community participation) [48]. The mediation test is used to test the significance level of the mediation effect. All models are robust and continuous variables are converted to standard values [49].

5. Results

5.1. Influencing Factors and Intergenerational Differences in Community Integration

In line with common practice, we define migrant workers born before 1980 as the “the first generation of ATP” and those born after that time as the “the second generation of ATP”. The results of the Oprobit regressions are shown in

Table 2. The regression results.

Variables	Indicators	Panel A: the Entire Sample	Panel B1: the First Generation of ATP	Panel B2: the Second Generation of ATP
Personal characteristics	Gender	0.043 (0.027)	0.027 (0.038)	0.072 * (0.038)
	Age	−0.002 (0.002)	−0.004 * (0.003)	0.007 (0.006)
	Education level	0.063 *** (0.017)	0.048 * (0.030)	0.064 *** (0.021)
	Number of household members	0.008 (0.126)	0.031 (0.019)	−0.024 (0.018)
	Marriage status	0.117 ** (0.039)	0.029 (0.075)	0.144 ** (0.055)
Community Identity	Self-identity	0.318 *** (0.020)	0.312 *** (0.028)	0.323 *** (0.028)
	Local Acceptance	1.015 *** (0.025)	1.006 *** (0.035)	1.031 *** (0.037)
	Organizational identity	0.024 ** (0.011)	0.019 * (0.015)	0.032 * (0.016)
	Social identity	0.049 ** (0.018)	0.047 * (0.025)	0.045 * (0.026)
Community participation	Making suggestions	0.564 *** (0.049)	0.486 *** (0.070)	0.636 *** (0.068)
	Participating in discussions	0.255 *** (0.045)	0.338 *** (0.069)	0.194 ** (0.059)
	Participating in activities	0.148 *** (0.026)	0.188 *** (0.037)	0.100 ** (0.037)
	Community concern	0.761 *** (0.023)	0.798 *** (0.033)	0.721 *** (0.034)
Pseudo R2		0.0319	0.3385	0.3269
LR Chi2		6115.76 ***	3162.94 ***	2962.37 ***
Number of observations		10,015	5021	4994

Note: Each column lists parameter estimates. Robust standard errors are in parentheses. ***, **, * represent significant at 1%, 5%, 10% levels respectively.

for the entire sample (n = 10,015).

Overall, the model performs good fitting results with significance at 1% level. The significance status of each variable in different panel data remained largely unchanged, reflecting the robustness of the model.

From the regression coefficients in Table 2, the indicator of education level among the control variables has a significant positive effect on community integration in the full sample, and it is positively correlated at the 1% level, indicating that those with higher educational levels are more inclined to integrate into the community. The indicator of marriage status is significantly and positively correlated with the community integration at the 5% level, suggesting that more stable marital status such as being married, positively influences community integration. Among the series of variables on community identity, the indicators of local acceptance and self-identity both significantly and positively influence the intention of community integration at the 1% level, while the indicators of organizational identity and social identity are slightly less significant, both positively influencing the community integration at the 5% level. This suggests that community identity is an important variable influencing community integration, with identification from self and local people having a particularly significant impact on the willingness to integrate into the community. Other influencing indicators include organizational identity due to having a local circle of contacts and social identity due to having applied for relevant social security. Among the series of variables on community participation, making suggestions, participating in discussions, engaging in activities and community concern are all positively significant at the 1% statistical level, indicating that these indicators all positively influence the community integration of ATP.

Analysing the intergenerational differences in the sample, the age indicator of personal characteristics passes the significance test at the 10% level within the sample of the first generation of ATP. Education level is significantly positively correlated at the 10% and 1% levels within the sample of the first and second generation of ATP respectively, indicating that this indicator still maintains a significant positive impact on community integration within the sub-sample, and that the positive impact of higher education on community integration is more significant in the second-generation group.

In the sample of the second generation of ATP, the gender indicator is positively significant at the 10% statistical level and the marriage status indicator is positively significant at the 5% level, but these 2 indicators are not significant in the sample of the first-generation group, showing intergenerational differences. Among the series of variables of community identity, the positive significant levels of the indicators of local acceptance and self-identity are consistent with the overall sample, while the indicators of organizational identity and social identity positively affect community integration at the 10% level and do not show differences between generations. This indicates that self-identity, local acceptance, organizational identity and social identity all have a significant positive effect on community integration among the first generation and the second generation of ATP. Among the series of community participation variables, the 2 indicators of suggestion and community concern are positively significant at the 1% level, consistent with the overall significance and do not differ between generations. The 2 indicators of participation in discussions and activities are both significant at the 1% level and consistent with the overall group, but the significance of these 2 indicators decreases in the second-generation group (from 1% to 5%). This indicates the positive effect of making suggestions and community concern on the community integration of both generations of the migrant population, but the significance of the positive effect of participation in discussion and participation in activities decreases among the second generation.

5.2. Model Robustness Tests

5.2.1. The Regression Result of Changing the Dependent Variable to “Household Relocation Willingness”

The robustness test examines the interpretation of the evaluation methods and indicators, that is, whether the evaluation methods and indicators still maintain a relatively consistent and stable interpretation of the evaluation results when some parameters are changed [50]. To test the reliability of the findings, the willingness of the ATP to (or to have) move their household to the local area is used as the replaced dependent variable [51]. The dummy variables are constructed by taking the values of “unwilling = 1, have not considered it = 2 and willing = 3”, and the main findings are tested for robustness [52]. The results of the tests are shown in

Table 3. Robustness analysis of regression results.

Variables	Indicators	Alternative to Variables			Alternative Regression Methods
		Panel A: the Entire Sample	Panel B1: the First Generation of ATP	Panel B2: the Second Generation of ATP	Panel A: the Entire Sample	Panel B1: the First Generation of ATP	Panel B2: the Second Generation of ATP
Personal characteristics	Gender	0.005 (0.023)	0.026 (0.033)	0.045 (0.033)	0.073 (0.050)	0.065 (0.071)	0.111 (0.072)
	Age	−0.010 ** (0.001)	−0.012 ** (0.002)	−0.004 (0.005)	−0.004 (0.003)	−0.009 * (0.005)	0.015 (0.011)
	Education level	0.132 *** (0.015)	0.189 ** (0.026)	0.096 *** (0.018)	0.107 ** (0.032)	0.090 ** (0.056)	0.106 *** (0.040)
	Number of household members	0.005 (0.011)	0.004 (0.017)	0.009 (0.016)	0.010 (0.024)	0.041 (0.035)	0.042 (0.034)
	Marriage status	0.141 *** (0.033)	0.130 * (0.641)	0.104 ** (0.047)	0.209 ** (0.072)	0.011 * (0.140)	0.246 ** (0.102)
Community identity	Self identity,	0.022 ** (0.010)	0.031 * (0.013)	0.010 * (0.014)	0.044 * (0.020)	0.036 * (0.028)	0.060 * (0.030)
	Local acceptance	0.170 *** (0.023)	0.213 *** (0.032)	0.123 *** (0.032)	1.965 *** (0.050)	1.957 *** (0.070)	1.985 *** (0.071)
	Organizational identity	0.152 ** (0.017)	0.145 *** (0.025)	0.161 *** (0.024)	0.563 *** (0.038)	0.554 *** (0.054)	0.574 *** (0.053)
	Social identity	0.035 * (0.016)	0.021 * (0.022	0.076 ** (0.023)	0.090 ** (0.034)	0.076 * (0.047)	0.093 * (0.049)
Community participation	Making suggestions	0.237 *** (0.045)	0.107 ** (0.064)	0.365 *** (0.064)	1.002 *** (0.094)	0.846 *** (0.132)	1.163 *** (0.135)
	Participating in discussions	0.095 ** (0.041)	0.240 *** (0.065)	0.150 ** (0.0540	0.458 *** (0.086)	0.613 *** (0.134)	0.346 ** (0.114)
	Participating in activities	0.082 *** (0.022)	0.035 * (0.032)	0.127 ** (0.032)	0.270 *** (0.048)	0.319 *** (0.069)	0.206 ** (0.069)
	Community concern	0.180 *** (0.022)	0.147 *** (0.032)	0.212 *** (0.031)	1.473 *** (0.046)	1.534 *** (0.065)	1.405 *** (0.066)
Pseudo R2		0.3560	0.3630	0.3531	0.3529	0.3596	0.3476
LR Chi2		2756.01 ***	1968.4 ***	1812.7 ***	1503.17 ***	1360.06 ***	1149.87 ***
Number of observations		10,015	5021	4994	10,015	5021	4994

Note:(1): In Table 3, the results in left column are the regression results when the dependent variable is replaced by “household relocation willingness”, and the results in the right column are the results when the regression method is replaced by Ologit model. (2): Each column lists parameter estimates. Robust standard errors are in parentheses. ***, **, * represent significant at 1%, 5%, 10% levels respectively.

. As can be seen from the replacement variable column of Table 3, the indicators in the community identity, community participation and control variables all have positive effects on the household migration variable at different levels of significance, and the effects differ between generations. The regression results remain consistent with that of the main regression. To save space, the results of the variable regressions are omitted from the analysis

5.2.2. Replace the Regression Method with Ordered Logit Regression Results

To test the reliability of the findings, the regression method is replaced with the ordered logit (Ologit) model to capture the characteristics of the variables [52]. The Ologit model obeys the logistic distribution, while the Oprobit model follows a normal distribution. Both models are commonly used models of discrete choice models. But the logit model is simple and direct, and it is more widely used. Therefore, we think that Ologit is suitable for the robustness test [53]. The results are also shown in Table 3. From the alternative regression method column, we can see that the replacement results remain consistent with the results of the main regression, indicating that the main regression results are tested to be robust.

5.2.3. The Mediating Effect of Community Identity between Community Participation and Community Integration

The community identity index and the community participation index are calculated by extracting one component for each of all items of community participation and community identity using principal component analysis [54]. Multiple regression analysis is conducted using SPSS Process component, and the mediating effect of community identity between community participation and community inclusion is tested by the bias-corrected percentile Bootstrap method. Samples are repeated 5000 times and 95% confidence intervals are calculated [55].

The results of the tests in

Table 4. Results of the mediating effects test of community identity in the process of community participation influencing community integration.

Resulting Variables	Predictive Variables	Fitted Indicators		Sig
		R2	F Value	T Value	p Value
Community integration	Community participation	0.31	74.46 ***	5.39	0.0000 ***
Community identity	Community participation	0.27	203.9 ***	4.52	0.0000 ***
Community integration	Community identity	0.35	268.5 ***	3.21	0.0000 ***
	Community participation			6.68	0.0000 ***

Note: Each column lists parameter estimates. *** represents significant at 1% level.

show that community participation is a significant predictor of community integration (t = 5.39, p < 0.001) and community participation is a significant predictor of community identity (t = 4.52, p < 0.001). The direct predictive effect of community participation on community integration remains significant when community identity is included (t = 6.68, p < 0.001). In addition, the direct effect of community participation on community integration is 0.366, with a direct effect of 44.26% of the total effect ratio, and the Bootstrap 95% confidence interval does not contain 0 (see

Table 5. Decomposition of mediating effects of community identity.

Items	Effect	Boot SE	T Value	95% Confidence Intervals		Relative Effects%
				LLCI	ULCI
Total effect	0.827	0.062	2.729 ***	0.480	0.793	100%
Direct effect	0.366	0.050	6.678 ***	0.238	0.435	44.26%
Indirect effect	0.461	0.041	/	0.256	0.389	55.74%

Note: Each column lists parameter estimates. *** represents significant at 1% level.

), indicating that the direct effect of community participation in enhancing community integration is significant. The indirect effect of community identity in the effect of community participation on community integration is 0.461, with an indirect effect of 55.74% of the total effect ratio, and the Bootstrap 95% confidence interval does not contain 0. This indicates that community participation not only directly predicts community integration, but also predicts community integration through the mediating effect of community identity. The model of the mediating effect between variables is illustrated in Figure 2a,b are the mediating effects of the mediating variable community identity; c’ is the direct effect.

5.2.4. Survey on Willingness to Integrate and Reasons for Staying

The questionnaire is set up to investigate the reasons of ATP for staying. After eliminating invalid questionnaires in cases such as this item not being filled in, the number of valid questionnaires returns for this question is 7799.

Table 6. Reasons for staying of the ATP with a desire to integrate.

Reasons	n	%
Better personal development	2640	36.5
Better educational opportunities for the children	1670	23.1
Access to higher income	1450	20.0
Access to public services such as transport and health	1005	13.9
Local social network	245	3.4
Other	224	3.1
Total	7234	100

shows the descriptive statistics of the valid questionnaires.

As can be seen from Table 6, most respondents stay in URICs for better development of individual children or families (36.5%, 23.1% and 20.0% respectively). Among them, 13.9% say they want to access better public services in the city, and 3.4% say they stayed because they had built up a social network in the local area. The survey results suggest that some of the ATP stay in the community because of their personal or family members’ future development, such as access to job opportunities, education resources and higher income. Some of the reasons for staying in the community are access to public services and the establishment of a social network.

6. Discussion

We must acknowledge that China’s urban-rural divide over the past half-century has resulted in an uneven development between urban and rural areas. Unequal distribution of resources has led to a large gap between urban and rural areas in education, infrastructure, social welfare, and other aspects [56]. In the process of urbanization, there is even a status inequality between residents with a local household registration and ATP without a local household registration. Immigrants who are temporary residents are not treated as city residents and have few rights in terms of labour and employment, housing, health services, education, and welfare guarantees [57]. To reach urbanization of the city, the needs of ATP in these areas must be taken into account to achieve the urban integration and community integration of the ATP.

In terms of social sustainability, poverty and social exclusion are the most worrisome issues. In our study, ATP’s self-identification has a greater impact on community integration. In addition, the effects of including organizational identity due to having a local circle of contacts and social identity due to the relevant social security are also significant. This is consistent with the results of previous studies. Also, community identity plays a mediating role in community participation and community integration. Therefore, it is necessary to strengthen the community identity of the migrant. Krzysztofik et al. illustrate the problem from the opposite side in their study. The marginalized encounters of the migrant population in the city make it difficult for them to integrate into temporary settlements. The situation is exacerbated by the deceptive assistance of the local government in solving social problems, which can even lead to negative outcomes such as loss of self-esteem, depression, and crime among the migrant population [58]. In the post-military estate of Rolnicza in Europe, the problem of community integration of migrant residents can be solved by demolishing the existing houses and rebuilding them. This may be an idea for solving the community integration difficulty. But we must realize that a lot of demolition and reconstruction in China may not be a feasible solution considering the actual local situation.

The results of the model show that there are varying degrees of intergenerational differences among the factors influencing community integration. This is supported by the study of OUŘEDNÍČEK et al. [59]. They suggest that the solution to the problem of the migrant population should focus on the age structure. It is important to note that in China, following the great efforts to provide universal education, young people are generally more educated than the older, and with that comes greater opportunities and skills for community participation. Such a situation is also found in Poland [60]. Migrant young people play a leading role in urban fringe areas, which is an important practice to facilitate the integration of the migrant population. ATP of working-age indirectly contribute to economic activity in the city and help to promote further social and technological development. All these processes together contribute to the civilization process of ATP.

Due to the limitation of data availability, we only got the data of CMDS in 2017, and there is a certain lag in the data, which has an impact on the guiding effect of the relevant recommendations. The data are to be further updated in future studies. Secondly, the establishment of evaluation indicators for community integration, community identity and community participation is limited by the content of the questionnaire, and the indicator selection in the model is therefore limited, so the content of evaluation indicators needs to be further supplemented in the future research.

7. Conclusions and Recommendations

7.1. Conclusions

This paper use data from CMDS conducted by the National Health Commission of China in 2017. Based on the Oprobit model, we investigate the influence of community participation and community identity on community integration among ATP living in URICs, and try to explore the mediating effect of community identity in the process of community participation affecting community integration.

The main findings are that: (1) Overall, community identity and community participation of the ATP positively influence community integration, while other factors such as education level and marriage status also significantly influence community integration. Measures to strengthen community identity and community participation are necessary to promote community integration, while we should also pay attention to individual characteristics and taking appropriate measures. (2) There are intergenerational differences in these effects. The positive effect of higher education on community integration is more pronounced among the second generation of ATP, and there are significant intergenerational differences in indicators such as participation in discussions, participation in activities, gender and marriage status. The second-generation ATP can play a pioneering role to lead the community integration of the ATP group. (3) Community identity plays an important mediating role in the process of community participation influencing community integration. The majority of respondents’ reasons for staying in the city are motivated by the need for better personal, child or family development, the desire to access better public services in the city, and the establishment of a social network in the local area. The development opportunities of community integration should be used to further promote the citizenship process of ATP.

7.2. Recommendations

In the light of the above analysis, it is necessary to pay attention to the actual needs of the migrant population when formulating urban renewal policies. Make good use of the community as a spatial carrier, and give full play to the role of community integration in promoting the urban integration of the ATP.

(1)

Equalize the supply of public services for the ATP.

The government should respect the desire of some of ATP to stay in the city. On the one hand, increase financial investment to improve the current building conditions, sanitary environment, road traffic and communication facilities in URICs. On the other hand, adjust the policy of applying for subsidized housing and increase the supply of public rental housing. Set aside a certain proportion of welfare housing to solve the accommodation problems of the migrant population and their family members, so that they can settle in the city. Also, urban governments should uphold the principle of basic public service equalization, and accelerate the supply capacity of basic public services in cities. The content of basic public services should be gradually expanded following the core rights demanded by the citizenship of ATP. Guarantee the necessities of their life, focus on improving education, health and cultural services in URICs, and provide ATP with a variety of services such as medical care, employment, and human rights protection to make them willing to live down in the city.

(2)

Improve the community governance system with community autonomy as the mainstay.

Guide the ATP to achieve self-development and individual empowerment in the community, increase community identity, and make them integrate more quickly into urban society. The second generation of ATP often has higher education levels and the potential for community leadership. They have a certain sense of social participation and show interest in public affairs such as national policies and social development. They can be encouraged to play a bonding role and participate in community public management, thus stimulating the initiative and enthusiasm of the ATP group to participate in community governance. Transform the community management model, make full use of new media to enhance the transparency of information, and guide community members to fully express themselves in an orderly manner based on mobile spaces, to achieve active participation and integration. Community neighborhood committees can be used to carry out various types of community activities, to promote community interaction and strengthen the perceived identity of ATP as members of the community, thereby enhancing their sense of belonging and identification with the community and the city.

(3)

Promoting ATP citizenship with community integration.

The essence of community integration is to realize collaborative governance between the migrant population and residents. Promote subjective motivation of the ATP to expand their social circle through community activities. Through community participation, ATP can strengthen the sense of common values in the community, increase interaction and form community identity, thereby restoring the spirit of altruism and public service in the traditional value system and strengthening the cohesion of community members. At the same time, strengthen employment guidance for migrant workers, improve their vocational skills, and enhance their competitiveness in employment. The government should actively promote employment services, timely release information on labour demand, and guide labour training. Thus, they can accumulate the ability to integrate into the city sustainably.

The existence of URIC makes cities more inclusive and allows the migrant population to live and work in cities, build a living base and establish a proper social status. The acceleration of urbanization in China depends to a large extent on the urbanization of the migrant population. We should face up to the existence of URIC and use the transformation of URIC as a breakthrough to promote the integration of the ATP into the cities, to reduce the cost of citizenship, thereby accelerating the process of urbanization in China and enabling all classes to benefit to the maximum from the accumulation effect of urbanization.