Background

Globally, migration from rural-to-urban areas is linked to mental health issues in China (Yang et al. 2018), deaths from air pollution in Peru (Carrasco-Escobar et al. 2020), depression in Ethiopia (Erulkar and Medhin, 2022), and migration hazards in Senegal (Boujija et al. 2022). At the moment, scholars are paying attention to the connection between migration and subjective well-being (SWB). Negative outcomes for parents who are left behind (Ghimire et al. 2018; Yahirun and Arenas, 2018; Vanore et al. 2018; Kumar, 2021; Teerawichitchainan and Low, 2021), children who are left behind (Zhao et al. 2018a; Gassmann et al. 2018; Antia et al. 2020; Chen and Zhou, 2021; Kharel et al. 2021), and left-behind spouses (Bhargava and Tan, 2018) are all linked to disruptions to familial norms of coresidence. Another study (Gebeyaw et al. 2022) shows that older homeless rural-urban migrants (RUMs) in Ethiopia have a high prevalence of poor well-being. Furthermore, it is frequently documented that institutional factors—rather than personal traits and living circumstances—play a larger role in the migration patterns of RUMs in China.

China has been using the Hukou (household registration) system since 1958 in order to regulate social protection, maintain social stability, and regulate internal migration between rural and urban areas. The Hukou system was loosened and deregulated until the 1980s, which gave rural laborers the chance to work in urban areas. Since 1992 and 2000, respectively, there has been a sharp rise in the number of first-generation RUMs born before 1980 and new-generation RUMs born in or after 1980 in urban China (Chen and Wang, 2015). According to Schoolman and Ma (2012), first-generation RUMs primarily work in “dirty and hard” industries and migrate temporarily (Zhu, 2007). Chinese policy reforms reduce labor market fragmentation and increase employment opportunities for first-generation RUMs, which is a migration incentive (Seeborg et al. 2000). First-generation RUMs appear to be at a disadvantage due to endowment effects (Messinis, 2013) and the rural-urban income gap (Zhu, 2002; Zhang and Song, 2003). Essentially, an economic theory can account for a portion of the SWB of first-generation RUMs (Harris and Todaro, 1970). Finding the causes of their declining SWB can aid in stabilizing and determining how best to help them.

Compared to first-generation RUMs, new-generation RUMs are more educated and skilled, and likely to work in manufacturing and service industries rather than the construction industry (Zhao et al. 2018b). In contrast to their predecessors, new-generation RUMs have better human capital, a strong desire to settle in cities, and a lower regard for Hukou benefits (Tang and Feng, 2015; Huang et al. 2017). In Chinese cities, registered individuals may be eligible for access to government services like health insurance, property rights, and public education programs. According to empirical evidence, RUMs’ conversion to urban Hukou status is negatively impacted by socioeconomic inequality in urban labor markets (Wu and Zheng, 2018).

However, in contrast to urban residents, RUMs continue to experience widespread discrimination based on Hukou (Meng and Zhang, 2001; Lu et al. 2013; Cheng et al. 2013; Zhang et al. 2016; Zhu, 2016). Hukou can determine a person’s points of access to social services, like hospitals and schools. Workers who migrate within urban China without changing their Hukou are regarded as second-class citizens. As a result, severe discrimination against rural migrants is a common practice among employers, governments, and urban residents. This discrimination has a substantial impact on RUMs’ psychological distress and quality of life (Wang et al. 2010). Rural Hukou continues to be a significant barrier to the institutional mobility of older rural migrants (Cheng et al. 2019). However, according to Xu and Palmer (2011), RUMs’ networks do not seem to have a significant impact on their level of political participation or life satisfaction. The intention of an urban settlement is significantly influenced by welfare programs (Sun et al. 2022) and urban amenities (Liao and Wang, 2019). From the standpoint of institutional design, the conversion of rural areas into urban Hukou has a somewhat protective effect on the psychological health of today’s RUMs (Song and Smith, 2021). Thus, people move to and live in cities as their psychological well-being improves.

At the moment, the majority of RUMs are concentrated in urban areas with underprivileged neighborhoods and subpar residential environments. Large cities are highly preferred by RUMs in China (Xing and Zhang, 2017), particularly megacities (Mohabir et al. 2017; Song et al. 2022). Rarely do RUMs construct their own homes; instead, they either live in housing provided by their employers or rent rooms from long-term residents (Mobrand, 2006). Similarly, China’s “urban diseases” are brought on by the RUMs due to their large number, high concentration, marginalized social standing, and unjust treatment (Gu et al. 2007). Economic incentives play a significant role in the settlement intention of RUMs belonging to the new generation (Chen and Liu, 2016; Chen and Wang, 2019). The housing circumstances of the three categories of new-generation RUMs (labor RUMs, intellectual RUMs, and entrepreneurial RUMs) vary and are mostly influenced by socioeconomic variables and Hukou status (Li, 2010). In eight Chinese cities, social connection is also said to affect the RUMs’ feeling of place in their host cities (Huang et al. 2019). This motivates us to look into the potential influence of socioeconomic factors on residential well-being (RWB) and sense of community (SoC) in Chinese RUMs.

It is crucial to conduct research on SoC and RWB among RUMs in China. First, they don’t receive well-being promotion from urban residents. According to a Shanghai study, among RUMs in urban China, the quantity of local ties is not a significant predictor of good mental health (Jin et al. 2012). Secondly, their promotion and well-being are unaffected by institutional reform. A thorough review of the literature demonstrates that the Hukou system as it exists now negatively affects China’s rural-to-urban migration (Song, 2014). According to a recent study by Chen and Zhao (2017), expanding urban Hukou entitlement won’t in and of itself encourage more rural-to-urban migration. RUMs therefore depend on the institutional components of SoC and RWB. However, no research has been done up to now that details the institutional components of SoC and RWB and how they relate to each other in Chinese RUMs. This research endeavors to address the void by utilizing publicly accessible survey data.

This paper aims to investigate the factors related to the SoC and RWB of Chinese RUMs. This is how the remainder of the paper is structured. Basic concepts, influencing factors, theoretical underpinnings, and research questions are covered in the literature review section. The data source, variables of interest, estimation framework, and statistical strategies are introduced in the method section. The primary findings of this work are shown in the following section. The last section comes to conclude.

Literature review

According to Hoffmann et al. (2019), the push-pull paradigm can be extended to consider the attitudes and choices of RUMs as drivers of migration in India. Theoretically, a couple’s decision to migrate is determined by their SWB (Tam and Grimes, 2022). According to a recent study, RUMs in China may have better psychological health if they have a sense of urban identity (Li et al. 2022a). Therefore, opinions about one’s neighborhood and community may have an impact on migrants’ psychological health in China. Globally, factors that influence migrants’ psychological well-being include age at migration in Surinamese background (Guo et al. 2019), stressors in Sweden (Solberg et al. 2021), health behaviors in the United States (Jagroep et al. 2022), and job mobility in China (Zhang et al. 2022). Individual socioeconomic characteristics are also probably related to RUMs’ well-being.

Professionals in social work, planning, and community development, as well as social theorists and researchers, have turned their focus to SoC and RWB. Although they are frequently disregarded, they are both significant aspects of SWB. Over the course of more than 25 years, Diener has published 200 works on the subject of SWB, or people’s positive emotions and life satisfaction (Diener, 2009). Diener’s (1984) review gives theories emphasizing psychological factors a lot of weight. According to Diener and his allies, SWB, or what is commonly referred to as “happiness”, is made up of several distinct but somewhat related variables (Diener et al. 2009). According to Diener and Suh (1997), social indicators, representative economic indicators, and SWB measures are required to assess a society. In the meantime, family and individual factors strongly predict life satisfaction (Morgan et al. 2011). Furthermore, an empirical study demonstrates the strong correlation between happiness and meaningfulness and family and social relationships (Delle Fave et al. 2011). Previous research indicates that SWB at the community level seems to influence that at the individual level as well (Ng and Fisher, 2016). Within a given city, RUMs may have different SoC and RWB.

SoC’s definition and influencing factors

SoC is the subject of a significant amount of research. “An acknowledged interdependence with others, a willingness to maintain this interdependence by giving to or doing for others what one expects from them,” is how Sarason (1974) first conceptualized SoC. Four components makeup SoC, according to McMillan and Chavis (1986): membership, influence, integration and need fulfillment, and shared emotional connection. Similarly, Royal and Rossi (1996) offer a fresh conceptual framework for comprehending SoC in educational and professional settings. As a result, SoC is used as a standard to evaluate how social and geographic communities are developing (Pretty et al. 2007).

According to Ditchman et al. (2017), the SoC dimension reinforcement of needs is a significant independent predictor of life satisfaction. Similarly, a number of academics discuss how community psychology can benefit from a methodical examination of the creative tension that exists between diversity and SoC (Townley et al. 2011). Furthermore, diversity and community are, in fact, bound together by the core principles of community psychology (Brodsky, 2017). Social support and a sense of community have become increasingly important in response to urbanization (Bess et al. 2002). Adolescent SoC concurrently exhibits strong psychometric qualities (Cicognani et al. 2012). Furthermore, promoting positive outcomes in multicultural communities has practical implications for the operation of multiple psychological systems of care (Brodsky and Marx, 2001; Brodsky, 2009). It’s possible that SoC and RWB among RUMs differ in China.

Theoretically, SoC can be predicted by social support from positive affect (Wang et al. 2015), considerably improve social well-being (Cicognani et al. 2015), and have a catalytic effect on block association participation (Chavis and Wandersman, 1990). A different study indicates that favorable results may be linked to a lack of psychological SoC (Brodsky, 1996). Furthermore, higher SoC is predicted by the frequency of participating in neighboring behavior (Farrell et al. 2004) and linked to formal group involvement (Albanesi et al. 2007). Similarly, an empirical study shows that SoC has a positive correlation with quality of life and social support but a negative correlation with daily hassles (Mak et al. 2009). According to recent data, SoC is positively correlated with the frequency of ritual participation and strongly linked to increased collective efficacy (Perkins and Long, 2002; Sohi et al. 2017). It’s possible that China’s community neighborhood conditions have similar effects.

Furthermore, it has been shown that socioeconomic factors—primarily in Western nations—are linked to SoC. Age profiles, for instance, show a distinct pattern of psychological SWB (Ryff, 1989). By ethnic identity group, analyses show a significant difference in SoC and positive affect (Kenyon and Carter, 2011). Empirically, SoC is associated with quality of life (Gattino et al. 2013), social integration (Prati et al. 2016), life satisfaction (Prezza and Costantini, 1998; Hombrados-Mendieta et al. 2013; Moscato et al. 2014; Zhang et al. 2017), SWB (Davidson and Cotter, 1991; Chi et al. 2017), and community participation (Ramos et al. 2017). According to a Prezza et al. (2001) study, neighborhood relationships, years of residence, marital status, group participation, and area of residence are the factors that predict SoC. Similarly, income has little bearing on SoC (Jorgensen et al. 2010). According to a different study conducted in Baltimore City, low-income communities’ differences in individual and community-level characteristics contribute to the heterogeneity for psychological SoC (Brodsky et al. 1999). There are currently no pertinent reports available in China, particularly among Chinese RUMs.

RWB’s definition and influencing factors

In the 1960s, 1970s, and early 1980s in the United States, attention to RWB corresponds with attention to “residential satisfaction” in the domains of urban affairs, urban planning, environmental design, and urban and community sociology. According to Sirgy (2016), RWB is currently defined as living arrangement satisfaction to the extent that this satisfaction enhances one’s sense of well-being in key life domains. Furthermore, a study carried out in a socioeconomically disadvantaged region demonstrates that children’s perceptions of residential quality are significantly influenced by the residential context (Cicognani et al. 2008a). Similarly, in stressful communities, Dupéré and Perkins (2007) propose that social isolation from neighbors may be protective to mental health. Furthermore, high levels of social connectedness among neighborhood residents are linked to the availability of activities and gathering spots (Lenzi et al. 2013).

The connection between place of residence and psychological well-being has been extensively studied in the literature. For instance, research conducted in the Columbus, Ohio, metropolitan area comes to the conclusion that the correlation between psychological well-being and residential satisfaction is a product of both variables’ reciprocal relationships with personal resources (Schwirian and Schwirian, 1993). Cicognani et al. (2008b) provide evidence in support of the positive correlation between SoC and social participation. A cross-sectional survey shows how the physical environment can improve frail older people’s social well-being (Nordin et al. 2017). According to a study, the residential environment has a greater influence on SWB in older adults in Shanghai than individual resources (Liu et al. 2017a). In particular, a Hong Kong study indicates that living circumstances affect older people’s psychological health and residential satisfaction (Phillips et al. 2005). Qian and Zhu’s (2014) study on RUMs’ sense of place in Guangzhou, China, reveals a notable distinction between RUMs’ sense of place in a city and a community. Furthermore, prior research has connected the SWB of RUMs to their living situations (Zaff and Devlin, 1998). Living conditions may therefore have some impact on SoC and RWB.

Theoretical bases

According to recent research, inhabitants with a higher sense of community are more likely to have higher levels of meaning in life (Kagan et al. 2023) and satisfaction with their community (Gamo and Park, 2024). The well-being of migrants rests on interpersonal relationships and interactions within a large society, as per Durkheim’s theory of social integration (Treviño, 2023). Following the publication of Jokl’s (1981) Theory of the Physical Environment, Kwon et al. (2019) report on the effects of the physical environment on health, happiness, and life satisfaction. Table 1 lists the main elements in SoC and RWB in accordance with recent references, taking into account the need for research on Chinese migrants.

Table 1 Main components in SoC and RWB in the references.

Basic research questions

Nevertheless, the effects of living arrangements and neighborhood conditions on SoC and RWB as reported by Chinese RUMs, have not been examined in any of these earlier studies. Consequently, the literature currently in publication highlights a significant vacuum in the empirical understanding of, for instance, the significance of housing or community-related issues in SoC and RWB. In particular, endogeneity, simultaneity, confounding, multicollinearity, and heteroskedasticity are frequently not taken into account at the same time in earlier research. As a result, the empirical research adds to the relevant associations. Thus, the following questions are the focus of the current study’s investigation:

  1. (a)

    Which are the primary socioeconomic factors that affect SoC and RWB, respectively, among living conditions and community neighborhood conditions?

  2. (b)

    Which are the primary socioeconomic factors—living arrangements, neighborhood conditions, and community—that affect SoC and RWB at the same time?

The independent variables in this study are living circumstances, community neighborhood conditions, and socioeconomic factors. It is possible to consider SoC and RWB as dependent variables separately or simultaneously. This study adds to the body of literature by investigating the interrelationship between SoC and RWB in addition to demonstrating the effects of independent variables on SoC and RWB.

Method

Data source

The survey “Development of Migrant Villages under China’s Rapid Urbanization: Implications for Poverty and Slum Policies (DMVCRU, https://reshare.ukdataservice.ac.uk/850682/)” was sponsored by Cardiff University’s Professor Wu (2015) and served as the basis for the analysis in this paper. The purpose of the survey is to look into how migrant villages form dynamically. In 2010, a face-to-face interview project was conducted in three major Chinese cities: Beijing, Shanghai, and Guangzhou. Twenty urban villages are chosen at random from the list of villages for each city. Twenty households are chosen at random using a random start address that is changed at predetermined intervals. There are 1208 valid questionnaires gathered in total.

Definitions of SoC score and RWB score

A module of the DMVCRU does exist that measures SoC with 12 items that reflect neighborhood and community perspectives and RWB with 14 items that measure housing and environmental status. In terms of internal consistency, SoC and RWB have Cronbach’s alpha coefficients of 0.8775 and 0.8150, respectively. According to the agreement measurement, each item accepts responses with scores ranging from 1 (very dissatisfied) to 5 (strongly agree) in SoC and RWB. The SoC score and RWB score, which range from 5 to 60 and 5 to 70, respectively, can therefore be determined by adding up all of their items. Consequently, a high SoC score may indicate a more regarded neighborhood and community atmosphere. A high RWB score may also indicate a more comfortable living space. The mean and standard deviation of each item, as well as descriptive statistics for the full text of the SoC and RWB items, are provided in Table 2.

Table 2 Descriptive statistics for the full text of SoC and RWB items.

Figures 1, 2 also show the histograms of the estimates to closely follow the shape of the normal density, respectively. After calculation, the Shapiro–Wilk W-test reports that the RWB score (n = 878, W = 0.99758, V = 1.358 p = 0.22531) and SoC score (n = 878, W = 0.99688, V = 1.746, p = 0.08498) are distributed normally, respectively. Shapiro–Francia W’ test reports that RWB score (n = 878, W’ = 0.99722, V’ = 1.657, p = 0.12533) and SoC score (n = 878, W’ = 0.99688, V’ = 1.859, p = 0.07930) are distributed normally, respectively. As for the RWB score, skewness and kurtosis test for normality reports probability values are 0.4447 for skewness and 0.3582 for kurtosis (joint p value = 0.4890), respectively. Simultaneously, the skewness and kurtosis test for normality report probability values are 0.2286 for skewness and 0.2808 for kurtosis (joint p value = 0.2710), respectively. Thus, the RWB score and SoC score are normally distributed on the angle of skewness and kurtosis. Thus, the SoC score and RWB score can be used as dependent variables for ordinary least-squares (OLS) regression analyses.

Fig. 1
figure 1

Distribution of SoC score.

Fig. 2
figure 2

Distribution of RWB score.

Independent variables

The socioeconomic profiles mainly include age (in years), gender (0 = female, 1 = male), marital status, educational attainment, registered residence, and household monthly income (Chinese Yuan). Marital status is grouped as single (0= unmarried and divorced or widowed) and married status (=1). Educational attainment is categorized as junior high school and below (0 = no education, primary school, and junior high school) and senior middle school and above (1 = senior middle school, technical secondary school, college, undergraduate, and postgraduate). In the sample, there are only four RUMs with postgraduate degrees. Thus, they are treated as missing values. The registered residence is divided into agricultural household (0 = city agricultural household and rural agricultural household) and non-agricultural household (1 = city non-agricultural household and rural non-agricultural household).

Living conditions include home ownership, number of bedrooms, number of living rooms, and number of housing facilities. Home ownership of living conditions is grouped by shared accommodation (=0) and own home (=1). Here, number of housing facilities is measured by the question: “Do you have the following housing facilities?” with response options of yes (=1) and no (=0). The items of housing facilities included a separate kitchen, separate toilet, shower facilities, liquefied gas, gas pipeline, air conditioning, heating equipment, and internet. Thus, the number of housing facilities is calculated by the sum of all of them.

Community neighborhood conditions include the number of problem-solving channels and converted residences. The number of problem-solving channels is measured by the question: “When the resident faces a problem regarding housing or community issues, would he/she solve the problems using the following channels?” with response options of yes (=1) and no (=0). The problem-solving channels include old friends and neighbors, neighborhood committee, village committee, property management company or developers, owners’ committee, relevant governmental departments, media, and others. Thus, the number of problem-solving channels is calculated by the sum of all of them. The variable of converted residences is measured by the question: “Do you think that the residences will be converted? ” with the response options of no (=0) and yes (=1).

Regression equations

The regressions for RWB score have the following form:

\({RWB}\,{score}=\beta 0+\beta 1\times {Age}+\beta 2\times {Gender}+\beta 3\times {Being\; married}+\beta 4\times {Educational\; attainment}+\beta 5\times {Hukou}+\beta 6\times {Ownership}+\beta 7\times {Family}\mbox{'}{s\; household\; income\; per\; month}+\beta 8\times {Number\; of\; housing\; facilities}+\beta 9\times {Number\; of\; bedrooms}+\beta 10\times {Number\; of\; living\; rooms}+\beta 11\times {Number\; of\; problem}-{solving\; channels}+\beta 12\times {Converted\; residences}+u1\)

The regressions for SoC score are defined analogously and have the following form:

SoC score = γ0 + γ1×Age + γ2 × Gender + γ3 × Being married + γ4 × Educational attainment + γ5 × Hukou + γ6 × Ownership + γ7 × Family’s household income per month + γ8 × Number of housing facilities + γ9 × Number of bedrooms + γ10 × Number of living rooms + γ11×Number of problem-solving channels + γ12 × Converted residences + u2

Here, β and γ are estimated parameters in the two regression equations, respectively. u1 and u2 are random error terms and are often assumed to be independent and identically distributed, respectively.

Statistical strategies

Due to the fact that the SoC score and RWB score are close to continuous measures, the suitability of ordinary least-squares (OLS) models is a potential analytical tool. First, in order to assess multicollinearity amongst the independent variables, the estat vif command (in the Stata program) is adopted to calculate the variance inflation factor (VIF) values for the independent variables after the regression. In the results from collinearity diagnostics, a group of independent variables with VIF values (<10) cannot be necessary to merit further investigation. Simultaneously, the multicollinearity is not a problem and can be safely ignored. Meanwhile, the Breusch–Pagan/Cook–Weisberg test for heteroskedasticity and Cameron and Trivedi’s decomposition of IM-test are employed to assess heteroskedasticity. If the p value in the tests is (preferably) 0.05 or smaller, then there is heteroskedasticity. Before heteroskedastic linear regression, the least-squares residuals against the value of the SoC score and RWB score are plotted by using the rvpplot command after OLS regression in order to model the variance function, respectively. After independent variables to model the variance are selected, heteroskedastic linear regressions are performed to explore the relationship between independent variables of interest and SoC score and RWB score, respectively.

Second, with respect to confounding effect, instrumental-variable quantile treatment effects (Stata program: ivqte) with quintile (0.5) is adopted to estimate the effect of Hukou status (home ownership) on SoC score and RWB score with a potential instrument (home ownership/Hukou status). The control variables are continuous variables (age, number of housing facilities, household monthly income, number of bedrooms, number of living rooms, and number of problem-solving channels) and dummies (gender, marital status, educational attainment, and converted residences).

Third, OLS linear regression with suppressing the constant term (Stata program: regress, noconstant) is adopted to analyze how living conditions and community neighborhood conditions influence SoC score and RWB score in the scenarios of shared accommodation, own home, and total sample, respectively. Regarding simultaneity, simultaneous equation models are applied when the phenomena are assumed to be reciprocally causal.

Thus, fourthly, three-stage estimation for simultaneous equations (Stata program: reg3) is adopted to analyze how living conditions and community neighborhood conditions influence SoC score and RWB score in the scenarios of shared accommodation, own home, and total sample, respectively. The method combines seemingly unrelated regression with two-stage least-squares estimation, which estimates all coefficients simultaneously and permits correlations of the unobserved disturbances across several equations. Finally, Zellner’s (1962, 1963) seemingly unrelated regression (Stata program: sureg), multivariate regression (Stata program: mvreg), and mixed-process regression (Stata program: cmp) are adopted to perform joint tests in order to correct endogeneity, respectively. Actually, correlation rho can reflect the interrelationship between the SoC score and the RWB score.

Results

Table 3 demonstrates that there are notable variations in home ownership based on factors such as RWB score, number of bedrooms, number of housing facilities, converted residences, gender and marital status, and household monthly income. The 900 heads of household in the study are split between two marital statuses (married: 74.49%, single: 25.51%), with men making up 84.53% of the participants. The age range of the sample is 15–80 years old (mean = 35.18 years; standard deviation = 12.54 years). 86.09% of the sample believes that the homes will be converted in the interim.

Table 3 Sample characteristics by home ownership.

In terms of residential status, a fraction of participants owns a home (14.06%) and shares accommodation (85.94%). Among the participants, 74.01%, or nearly three-fourths, have agricultural Hukou. He and Wang (2015) state that the new-generation RUMs thus predominate in the sample. A percentage of the participants (55.17%) states that they have completed junior high school or less in terms of education. In terms of employment, the sample (n = 898) is made up of: workers in state-owned economic units (3.90%); workers in urban collective economic units (4.68%); workers in other types of economic units (6.12%); owners of individual private enterprises in urban areas (32.29%); urban self-employed in private areas (33.30%); reemployed retired staff members (0.67%); workers in other sectors (9.58%); students (0.33%); homemakers (1.78%); retirees (3.90%); people awaiting allocation (1.34%); unemployed (0.67%); researchers (0.11%); and others (1.34%). Almost all of the participants live in two different housing facilities, have one bedroom, and no living room.

Figure 3 reports median RWB score in Guangzhou is the largest, followed by those in Shanghai and Beijing. The interquartile range of SoC score in Guangzhou is the largest in Fig. 4, with those in Beijing and Shanghai following closely behind.

Fig. 3
figure 3

RWB score in the three cities.

Fig. 4
figure 4

SoC score in the three cities.

Table 4 surprisingly demonstrates that VIFs in the linear regression on the RWB score and SoC score are less than 1.80. As a result, multicollinearity is safe to ignore and poses no threat. According to Cameron and Trivedi’s breakdown of IM-tests, there is strong evidence in favor of heteroskedasticity in the two regression equations.

Table 4 VIFs in the linear regression on SoC score and RWB score.

With the least-squares residuals against the value of SoC score and RWB score after OLS regression, age, number of housing facilities, household monthly income, and number of problem-solving channels are selected as the independent variables to model the variance in Table 5. Lnσ2 of the number of problem-solving channels significantly predicts the SoC score and RWB score positively, respectively. In a similar vein, the SoC score and RWB score are positively predicted by ownership, number of bedrooms, number of living rooms, and converted residences, respectively.

Table 5 Heteroskedastic linear regressions on SoC score and RWB score, coefficient (standardized error).

The statistical results of regressions using ownership and Hukou status as treatment and instrumental variables, respectively, are shown in Table 6. Hukou status has a significant impact on RWB score but not SoC score after adjusting for age, gender, marital status, educational attainment, monthly household income, number of living rooms, number of bedrooms, number of problem-solving channels, and number of converted residences. Ownership status has no discernible impact on the SoC score and RWB score after adjusting for age, gender, marital status, level of education, monthly household income, number of housing facilities, number of bedrooms, number of living rooms, number of problem-solving channels, and converted residences.

Table 6 Instrumental-variable quantile treatment effects on SoC score and RWB score, coefficient (standardized error).

In the shared accommodation sample, own home sample, and total sample, Table 7 demonstrates that the SoC score and RWB score are positively associated with age and the number of problem-solving channels, respectively. In the shared accommodation sample and total sample, respectively, the following variables have positively significant associations with SoC score and RWB score: gender, number of housing facilities, household monthly income, number of living rooms, and converted residences. In the entire sample, ownership has positively significant associations with both the SoC score and RWB score.

Table 7 OLS regression on SoC score and RWB score, coefficient (standardized error).

Table 8 demonstrates that in the shared accommodation sample, own home sample, and total sample, respectively, age, gender, educational attainment, number of living rooms, number of problem-solving channels, and converted residences have positively significant associations with SoC score and RWB score. However, in the three samples, the number of living rooms can result in a decrease in the SoC score and RWB score. In the meantime, the RWB score in the three samples may decrease due to Hukou status.

Table 8 Three-stage least-squares regression on SoC score and RWB score, coefficient (standardized error).

Table 9 demonstrates how the SoC score and RWB score in the three simultaneous equation models can be predicted, respectively, by age, gender, educational attainment, home ownership, number of housing facilities, household monthly income, number of living rooms, number of problem-solving channels, and converted residences. In the meantime, a decrease in the SoC score and RWB score can be caused by the Hukou status and the quantity of living rooms, respectively. In the seemingly unrelated regression, multivariate regression, and mixed-process regression, the estimated correlation coefficient is >0.56, respectively. The SoC score and RWB score thus have a moderate correlation.

Table 9 Simultaneous equation models on SoC score and RWB score, coefficient (standardized error).

Discussion

The SoC score and RWB score, respectively, are determined by the following factors: living conditions (home ownership, number of bedrooms, number of living rooms, and number of housing facilities), neighborhood conditions (number of problem-solving channels and converted residences), and sociodemographic characteristics (age, gender, marital status, and household monthly income). As far as the authors are aware, this is the first study to calculate the correlation between the SoC score and RWB score of RUMs who are temporarily registered, have low social and economic status, and have substandard housing. This study indicates a correlation between the SoC score and the RWB score. The study’s regression analysis reveals that, for the entire sample, the number of housing facilities, monthly household income, number of living rooms, number of problem-solving channels, and number of converted residences have significant but varying effects on the SoC score and RWB score.

Some published findings can help to explain the empirical results of this study in terms of socioeconomic factors. First off, this study’s results are consistent with those of Zeng et al. (2020). They discover that a number of sociodemographic factors, including gender, income, and education, as well as housing attributes like size and the standard of sanitary facilities and kitchens, significantly influence how satisfied new-generation RUMs are at home. According to this study, older RUMs have higher SoC score and RWB score than younger RUMs. An empirical study can clarify this. According to a Guangzhou study, RUMs are more likely to move into residential areas in the urban core as they get older (Li and Mao, 2018). Second, in China, gender affects RUMs’ SoC score and RWB score. A study conducted in the West found that gender differences account for individual happiness (Lera-López et al. 2018). Some empirical research can shed light on the gender role in happiness studies in the context of China. For instance, Fan et al. (2013) found that, out of all the groups, female RUMs experience the greatest barriers to healthcare access, based on data from 531 RUMs and 529 local urban residents in Shanghai, China, who are between the ages of 16 and 64. Gender discrimination serves as a further source of explanation. Jiang et al. (2011), for instance, discovered that gender discrimination has some detrimental effects on China’s development. Third, married individuals are more likely to have high SoC score and RWB score in the regression analysis. In comparison to single people, this indicates a decline in urban prejudice and an increase in the need for residential space. Fourth, the SoC score and RWB score are reflected in simultaneous equation models and are dependent on the monthly household income. Wage disparities between urban locals and RUMs can be used to support this. Because of the sectorial effect, wage effect, working time effect, and population effect, the annual earnings of urban residents in China’s cities are higher than the long-term earnings of RUMs (Démurger et al. 2009). Even though RUMs face institutionalized Hukou discrimination (Ouyang et al. 2017) and are paid a discrimination wage (Magnani and Zhu, 2012), they are still in competition for high levels of living conditions based on their SoC score and RWB score.

Additionally, it has been established that living conditions affect both the SoC score and RWB score. The SoC score and RWB score are positively correlated with home ownership. These are consistent with findings from several pertinent studies. Numerous studies show that RUMs’ access to affordable housing mechanisms in urban China is restricted by their Hukou status (Yu and Cai, 2013; Wu and Zhang, 2018). In the late 1990s, migrant households in urban China began to experience a new kind of housing poverty (Sato, 2006). It’s possible that migrant workers’ housing conditions won’t get better as a result of the relationship between the government and private providers (Li and Zhang, 2011). A study conducted in Jiangsu province on the housing tenure of China’s RUMs in eight destination municipalities reveals that Hukou, age, gender, educational attainment, household size, and personal income all have an impact on the prevalence of home ownership (Huang et al. 2014). This is also consistent with earlier studies showing that workers who get the accommodations they request are noticeably happier with them (Balser and Harris, 2008).

The study’s key finding is that a limited number of housing facilities have statistically significant positive correlations with SWB. Additionally, there is a negative significant correlation between the number of living rooms and SWB and a positive significant correlation between the number of bedrooms and SWB. The results can be explained by the fact that, as a result of their low income, the majority of RUMs reside in urban villages (Chengzhongcun, urban villages, or villages in the city). RUMs’ housing needs have been documented in a number of other studies (e.g., Wu, 2008a; Lin and Zhu, 2010). According to some Chinese studies (Lau and Chiu, 2013; Hao et al. 2011), urban villages can provide RUMs with the housing they need when economic necessity and policy constraints allow it (Xiang, 2000; Wang, 2016). Comparable to impoverished neighborhoods (Lin and Meulder, 2012), urban villages are distinguished by inexpensive housing associated with unclear property rights, an unofficial rental market, and a lack of state regulation (Liu et al. 2010). Beijing’s peri-urban area saw a concentration of thousands of youthful, highly educated RUMs after 2000, giving rise to a new type of urban slum (Zhao, 2013). A study using secondary data shows that the majority of RUMs, particularly in Beijing, struggle with housing overcrowding (Liu, 2017). RUMs are isolated from urban residents, according to studies conducted in Shanghai (Feng et al. 2006). They are concentrated in the peri-urban area, which is viewed as a potential crisis point that could jeopardize social and economic development (Wu, 2008b). According to research, RUMs living in urban China have lower life satisfaction levels (Cai and Wang, 2018) and an average happiness score (Knight and Gunatilaka, 2010) compared to rural households.

The statistical results of this study are consistent with certain findings from studies conducted in China with respect to community neighborhood conditions. For instance, a study suggests that urban Chinese migrants’ SWB can be improved by their satisfaction with the community environment (Lin and Huang, 2018). According to Bonnefond and Mabrouk’s (2019) investigation, internal migrants who have settled in cities but possess a rural hukou are less likely than their rural counterparts to report good life satisfaction. According to a different study, residents’ quality of life can be greatly enhanced by making physical community environments better from the perspectives of social, mental, and ecological well-being (Aman et al. 2022). Chinese adults who resided in communities with higher levels of community insecurity are likely to have lower levels of SWB, according to a 2018 study conducted in 160 communities in Yunnan, China (Wei et al. 2022). Neighborhood social support (Chen et al. 2021) and community support (Jia et al. 2021) can lower mental health risks and psychological distress during the COVID-19 pandemic.

This study’s findings about the correlation between SoC score and RWB score are consistent with those found in the Western world. For instance, a cross-sectional study conducted in the US suggests that SoC can both predict and affect the course of recovery for people residing in recovery homes (Stevens et al. 2018). However, this study contradicts another study that finds no significant correlation between SoC and neighborhood association or neighborhood physical characteristics (Kingston et al. 1999).

Based on earlier research, some empirical findings shed light on the redesign of policies. According to a number of recent studies, China’s urban housing reform efforts can be considered unsuccessful if they have a negative impact on the SoC score and RWB score. In actuality, RUMs’ social integration into cities is less complete than that of China’s urban-urban migrants (Qin et al. 2021). Among RUMs in China, home ownership may raise the possibility that they will use urban public health services (Wang et al. 2020). Realistically, several studies have discovered that the enhancement of RUMs’ housing is disregarded and overlooked by the present housing reforms in urban China (Shen, 2002; Jeong, 2011; Hu et al. 2014; Shen, 2015). Worse yet, a long-term study finds that RUMs face a significant “glass ceiling” in the urban labor market (Qu and Zhao, 2017). Empirically, the wages of migrant workers are statistically significantly impacted by a few neighborhood characteristics (Cheng and Wang, 2013). In particular, prior research indicates that the RUMs’ SWB varies in relation to community ties and work conditions (Akay et al. 2012). In particular, SoC in China has the ability to fully mediate the associations among neighborhood environment, meaning in life, and positive affect (Zhang and Zhang, 2018), predict local political participation (Xu et al. 2010), and mediate the relationship between public space and well-being (Zhang et al. 2018). In modern China, neighborhood interaction is regularly and significantly linked to community participation among RUMs (Palmer et al. 2011). Furthermore, when looking for social support, new-generation RUMs are probably going to turn to non-kin, non-territorial, and cross-class networks (Liu et al. 2012). Based on the aforementioned research findings, Chinese urban governments can therefore reconsider and redesign their housing policies for RUMs.

The main findings based on the historical data can be considered to provide reasonable implications for the present day, according to subsections of the Statistics on Children of Migrant Workers and Rural Left Behind Children in Schools in 2022 China Statistical Yearbook (http://www.stats.gov.cn/sj/ndsj/2022/indexeh.htm) and the Migrant population (100 million persons) in the 2020 China Statistical Yearbook (http://www.stats.gov.cn/sj/ndsj/2020/indexeh.htm). Thus, the primary conclusions drawn from the data in 2010 may be thought to have pertinent implications for the current situation in the 2020 s given that migration has increased since 2010. As a result, the empirical findings of this study offer policymakers fresh perspectives on how to enhance the SoC score and RWB score among RUMs. With an emphasis on Chinese urban administration, decision-makers can identify future areas where SoC score and RWB score can be enhanced by using the paper’s analytical process as a logic. Administrators who work with Chinese rural-urban migrants need to understand SoC score and RWB score. This could reduce administrative bias and enable the implementation of effective urban policies. From this vantage point, developing systems that improve the policy-design process requires identifying the factors that influence housing arrangements and neighborhood conditions. This can also aid in the process of designing policies and putting them into practice for environmental planning in migrant villages found in China’s urban areas.

Future research directions may be indicated by some important information that this study overlooked. The significance of social support in enhancing emotional well-being has been highlighted by certain studies conducted in Bangladesh (Koly et al. 2021), Sweden (Alexander et al. 2021), Western countries (Ekoh et al. 2022), and China (Zou et al. 2021). The RUMs, on the other hand, have lower levels of SWB than locals, and the social environment, social support, and feeling of relative deprivation all have an impact on the cognitive components of SWB (Liu et al. 2017b). Future studies can therefore compare the well-being of RUMs, urban residents, and peers from rural areas as a function of social support. Furthermore, because the questionnaire lacked important neighborhood variables, the study did not report on the role of neighborhood environments; however, other studies have documented this role (Schwanen and Wang, 2014; Ma et al. 2018). As a result, future study settings may include the neighborhood’s built, social, and natural environments.

Strengths and limitations

This study has two primary strengths. First, selecting the best strategies is aided by the provisional explanations of dependent variables. Second, to investigate the relationships of interest, the statistical design is based on multicollinearity, heteroskedasticity, simultaneity, confounding effect, and endogeneity. In addition, different applications of a range of regression models reflect the desired relationships objectively.

This study has two primary limitations. Firstly, the longitudinal associations of interest could not be represented by this cross-sectional study. The SWB of RUMs should be investigated further using panel data in order to identify generalized causal results. Secondly, it is impossible to conduct a comparative analysis without a sample of peers from rural and urban areas.

Conclusions

In conclusion, the present study’s findings indicate that living conditions and community neighborhood conditions influence the SoC score and RWB score among the RUMs. In addition, the SoC score has a moderate correlation with the RWB score. Moreover, institutional factors like Hukou also influence RUMs’ RWB score. The best way to increase well-being, according to this study, are to increase the number of converted homes and housing wealth. In the future, longitudinal data may be used to reflect the associations of interest through comparative and causative studies.