Differences in depressive symptoms between rural and urban Chinese labor force: The mediating effects of community factors

Abstract

Purpose:

This study aimed to examine differences in depressive symptoms between urban and rural workers in mainland China and to identify community factors that could contribute to such residential differences.

Methods:

This study used nationally representative data from the 2014 China’s Labor Force Dynamic Survey. Data were collected through face-to-face interviews on a sample of 22,073 participants from 29 provinces of China, including 15,098 rural workers (Mage = 44.92, standard deviation (SD) = 14.85) and 6,975 urban workers (Mage = 43.28, SD = 13.62). Mediators included community cohesion, foreseeable community threat, supportive network size and medical benefit coverage. Mediation analyses were conducted using Hayes’ SPSS Macro Process for multiple mediators.

Results:

Urban participants reported fewer depressive symptoms than their rural counterparts. Lower levels of community cohesion, higher community foreseeable threat and poorer medical coverage were related to fewer depressive symptoms. Rural–urban differences were mediated by community cohesion (B = −0.12, p < .01), foreseeable community threat (B = −0.08, p < .01) and medical benefit coverage (B = 0.25, p < .01).

Conclusion:

This study sheds light on distinctive roles of community factors in explaining rural–urban differences in depressive symptoms. Policies or programs should be designed to promote strengths and address weaknesses in rural communities.

Keywords

Rural–urban disparities depressive symptoms community cohesion community threat community supportive network size

The split Chinese rural and urban labor systems

Immediately following the establishment of the People’s Republic of China in 1949, China’s socioeconomic development was subject to the central government’s plan and control. The social structure featured a split urban–rural system characterized by a household registration system, urban state-owned factories that controlled supply and demand and rural communities that regulated the industry of farming (Wang, 2012). As a result, a split urban and rural labor system was formed, in which workers in urban and rural areas have differing employment benefits and welfare coverage. Overall, rural workers received inferior benefit coverage to their urban counterparts.

The social structure has undergone substantial changes since the more open economic policy was implemented in 1978, which shifted China’s economy from planning-based to market-oriented. The restrictions on population flow have been loosened. One associated phenomenon is the migration of working people from rural to urban areas, which before was highly restricted but now is common. Although income, access to housing and social security have been improved in rural areas, the gap between the rural and urban labor force remains wide and will continue in the near future. According to the National Bureau of Statistics, per capita income of urban residents in 2015 was about 2.73 times that of rural residents (Liu, 2016). The rural–urban differences in socioeconomic development and associated welfare benefits also have been argued to result in health disparities between rural and urban residents (Chi, 2014; Lowry & Xie, 2009).

Studies have consistently documented the differences of emotional well-being between rural and urban residents in China. Rural residents in general tend to have poor mental health, with anxiety and depression being the two most prominent psychiatric problems (Chen & Ni, 2006). Ding (2009) examined mental health status of 9,841 rural residents in an east coast region and found that more than one-fifth of them had clinically significant anxiety symptoms. Xue, Wang, and Lu (2005) found that among 16,640 residents in Hebei Province, urban workers including migrant workers had significantly better mental health than rural residents.

Known risk factors of depressive symptoms

Depressive symptoms are influenced by a number of factors. Enormous economic and work pressures to raise families, particularly for men, are likely to trigger psychological stress in workers (He, Keung, & Shouchui, 2010; Wong, He, Leung, Lau, & Chang, 2008). Socio-demographic factors, such as being female, unmarried or divorced and a member of ethnic minority groups, are associated with more likelihood of depressive symptoms (Cui et al., 2014). Lower self-rated social status and poorer physical health were consistent risk factors for depressive symptoms (Cole & Dendukuri, 2003; Gao & Xu, 2015).

While considerable research has been conducted to reveal individual-level risk factors associated with depressive symptoms among workers, little attention was paid to community factors. The most studied community-related factor is social support that one can obtain from relatives, friends or neighbors in the community. Social support reflects one’s social capital, which entails social resources and networks of relationships a person can rely on to deal with difficult situations (Berkman & Syme, 1979; He et al., 2010). Larger social support network size is a reliable predictor of lower depressive symptoms (Fiori, Antonucci, & Cortina, 2006).

Findings about another community factor – community cohesion – are limited and mixed. Some researchers (Kennelly, O’Shea, & Garvey, 2003; Subramanian, Kim, & Kawachi, 2002) argue that social participation and social trust in the community would not collectively strengthen individual health. Yet, other empirical evidence suggests otherwise. Snelgrove, Pikhart, and Stafford (2009) found that people living in areas with higher level social trust have better health outcomes, and a similar finding was observed in residents from four rural counties in three provinces in China (Qu, Wang, & Meng, 2008).

Research questions and hypotheses

Conclusion concerning these issues cannot be reached due to limited research on the effects of community factors on individual depressive symptoms. Therefore, this study aims to examine the effects of community-level factors on the emotional state of workers. Below, we described multiple community-level factors that we believe may influence workers’ emotional states and depicted the possible ways that these community factors influence emotional state.

As revealed by Kurt Lewin’s (1951) field theory, individuals are constantly influenced by their interaction with the surrounding environment – one’s daily living space. This exchange influences one’s mental health in two ways. One mechanism is via the spatial environment of personal life, including reconstructing personal living space and forming environmental structural factors. The other is to change the relationship in the daily living space, forming community-related factors that affect individual’s mental health. In other words, community factors can influence individual mental health by structural or relational embedding. Community embedded factors are factors that are community-based rather than individual-oriented, such as different community systems and community environments. In this article, we assume that due to the structural effects of long-term dual differentiation, rural and urban areas in China have marked differences in their natural environments, social environments and social systems, such as food safety, environmental pollution, infectious disease risk and unemployment. These variations in environmental pressures would affect rural and urban workforce differentially.

We further specifically hypothesize that rural and urban communities would differ in community cohesion, community threat, community support network size and medical insurance coverage, and that such differences would contribute to rural–urban worker differences in depressive symptoms. A conceptual model is provided in Figure 1. Below, we detailed our rationale for each of the four mediation hypotheses.

Figure 1.

A conceptual model of community factors mediating workers’ depressive symptoms.

Rural–urban differences in community cohesion would lead to differences in mental health between rural and urban laborers (Hypothesis 1). Rural and urban communities are different in the way residents interact with each other. Rural communities are groups based on kinship, ethics and unity, in which familiarity and mutual trust among villagers are higher than that in urban resident groups who tend to be more heterogeneous comprising strangers connected by division of labor.

Community threat reflects perceived risks (e.g. pollution and safety) in the community. Urban communities may face more threats due to industrialization and high concentration of population than rural communities. We hypothesize such rural–urban differences in community threats would contribute to their differences in depressive symptoms (Hypothesis 2).

Community supportive network size is the third factor that may affect differences in mental health between urban and rural areas (Hypothesis 3). An old Chinese saying goes, ‘a distant relative is no good as a close neighbor’. In both rural and urban areas, relationships based on informal interactions in neighborhoods are very important for individual’s mental health. People count on such informal networks for tangible and emotional assistance. Rural workers do not necessarily have larger network size than their urban counterparts (Zhao, 2008).

Finally, we hypothesize that differing healthcare coverage will mediate differences in depressive symptoms between rural and urban workers (Hypothesis 4). The current Chinese government–sponsored health insurance system includes five categories: new rural cooperative medical care, urban and rural resident medical insurance, urban resident medical insurance, urban worker medical insurance and governmental employee healthcare system. Ranked from highest to lowest, governmental employee healthcare system provides best coverage, followed by urban worker medical insurance, urban resident health insurance, urban and rural resident health insurance and new rural cooperative medical care (Tang, Wang, Qi, & Yu, 2014). The individual’s household registration and employment status determine his or her health insurance coverage. Studies have shown that older persons in the urban health insurance group reported better life satisfaction and subjective well-being than those in the rural cooperative medical group. In addition, their depressive symptoms were significantly fewer than those in the cooperative medical group (Tang et al., 2014).

Methods

This study used secondary data from the 2014 China’s Labor Force Dynamic Survey (CLFDS). CLFDS started in 2012 under the auspice of the Social Science Research Center at SUN-YETSEN University in China. Data were collected in 29 out of 31 provinces in mainland China, expect Tibet and Hainan provinces. Survey adopted a multistage and stratified sampling method. First, a sampling frame was created to include 2,282 districts and counties of the 29 provinces, and it was further strategized into six stratums by regions (east, middle and west) and population sizes (large and small). Counties were ranked by their gross domestic product (GDP) first and then population size. Within urban areas, street blocks were identified, and within each block, four urban neighborhoods were selected; in rural areas, townships were identified, and within each township, two villages were selected. Within each neighborhood and village, 35 households were randomly selected. Anyone aged 16 years or above who were employed at the time of interview was eligible. In 2014, the total completed survey sample was 23,594. This study excluded 1,434 individuals registered as rural household status but migrated to urban areas for employment, and 87 cases with significant missing values. The final sample size in this study was 22,073.

Table 1 presents demographic differences between rural workers (n = 15,098) and urban workers (n = 6,975). Rural workers reported higher depressive symptoms than their urban counterparts did. Rural workers tended to be older, receive less formal education, have poorer self-rated health, and lower self-rated class than their urban counterparts. However, rural workers reported higher job stratification than their urban workers. Rural workers had lower supportive network size and medical insurance coverage, but they reported higher community cohesion and less community threat than their urban counterparts did.

Table 1.

Sample characteristics by residential status.

	Rural workers (n = 15,098)	Urban workers (n = 6,975)
	M (SD)/percentage	M (SD)/percentage
Age**	44.92 (14.85)	43.28 (13.62)
Female	51.9%	52.0%
Married	87.0%	83.8%
Education**
No formal education	17.95%	2.40%
Elementary school	30.99%	8.32%
Middle school	35.01%	26.05%
High school	11.94%	29.26%
College and above	4.11%	33.97%
Self-rated health**	3.62 (1.03)	3.80 (0.90)
Self-rated class**	4.41 (1.64)	4.82 (1.70)
Self-perceived job satisfaction**	3.33 (0.82)	3.47 (0.75)
Community mediators
Community cohesion**	14.77 (2.47)	12.73 (2.54)
Foreseeable community threat**	9.20 (3.33)	10.28 (3.63)
Supportive network size**	5.75 (2.34)	6.08 (2.42)
Medical insurance coverage**	1.17 (0.62)	2.94 (1.49)
Outcome variable
Depressive symptoms**	9.43 (3.73)	9.06 (3.48)

SD: standard deviation.

Significance at .05 level.

Significance at .01 level.

Data collection

Data collection used computer-assisted personal interviewing strategy and spanned from June to October in 2014. Survey was delivered through face-to-face interviews and audio recorded. All data were uploaded directly to server of the research center in SUN-YETSEN University, where staff members managed data quality control, and data input, cleaning, storage and sharing.

Measures

Dependent variable

Depressive symptoms were assessed using a five-item scale, which was created by the research team and validated in Chinese populations (He, 2015). Participants were asked to rate how often in the past 4 weeks they felt sad, lost self-confidence, encountered insurmountable difficulties, were not able to get going and had work or daily life affected due to emotional problems. Possible responses ranged from 1 (‘never’) to 5 (‘always’). Higher scores indicated more depressive symptoms. Cronbach’s alpha in this sample was .856.

Independent variable

Those with an urban household registration status and worked in urban areas are considered urban workers (coded 0), and those with a rural household registration status and worked in rural areas were rural workers (coded 1).

Mediators

Community-level mediators consisted of community cohesion, foreseeable community threat, community supportive network size and medical coverage. Community cohesion was measured by a four-item scale that asked participants to rate to what degree (1) they are familiar with their neighbors and other residents in the community, (2) they trust their neighbors and other residents, (3) they help each other often and (4) the community is safe. The first three question responses ranged from 1 (‘not at all’) to 5 (‘very much’), and the last question ranged from 1 (‘not safe at all’) to 4 (‘very safe’). The total score ranged from 4 to 19, with higher scores indicating higher perceived levels of community cohesion. Cronbach’s alpha of community cohesion was .769 in this sample.

Foreseeable community threat was measured by six questions on a 4-point Likert-type scale. Participants were asked to rate the possibility from 1 (‘very unlikely’) to 4 (‘very likely’) regarding community threats they perceive in the next 5 years. The six questions consisted of likelihood of being unemployed, victimized by a criminal, attacked by a terrorist, affected by a contagious disease, subject to adulterated mediations or food and exposed to environmental pollution. The total score for these six items ranged from 4 to 24, with a Cronbach’s alpha of .834.

Community supportive network size was measured by asking about the number of people that they ‘feel close to’, ‘can ask for help’ and ‘can borrow at least $720 (about ¥5,000 Chinese) from’. The answers were open ended. Then, none was coded 1, ‘1–2’ coded 2, ‘3–5’ coded 3, ‘6–9’ coded 4 and ‘10 and above’ coded 5. Total scores were used, and the Cronbach’s alpha was .680.

Medical benefit coverage reflected the level of medical benefits participants had. Insurance types were ranked based upon the levels of benefits it entailed, with new rural medical cooperative insurance coded 1 and government employee medical insurance coded 5. Anyone who reported having employer subsidized insurance benefits or private insurance had their score increased by 1. If they had both, their level of medical coverage scores increased by 2. Therefore, the possible range of medical coverage was from 1 to 7, with higher scores indicating better benefits.

Control variables included sociodemographics, self-rated health, self-perceived social status and overall satisfaction with employment. Social demographics included gender, age and education levels. Self-rated health was measured by one-item that asked participants to rate their overall health condition on a Likert-type scale from 1 (‘very poor’) to 5 (‘excellent’). Self-rated social status was measured by asking participants to rank their perceived social class on a Likert-type scale from 1 (‘lowest’) to 10 (‘highest’). Perceived job satisfaction was measured by one-item asking participants to rate their overall satisfaction with their job on a 5-point Likert-type scale from 1 (‘not satisfied at all’) to 5 (‘very satisfied’).

Analysis strategies

T-tests were performed to reveal differences in sample characteristics and variables of interest between rural and urban workers; correlation analyses were run among the studied variables. Finally, we adopted Hayes’ Process Procedure for SPSS to perform the multiple mediator analyses. Controlling for demographics, self-rated health, self-rated class and perceived job satisfaction, we tested the main effect of rural–urban residence on depression, as well as its indirect effects through four possible community mediators. The Process procedure used a bootstrapping method to estimate the indirect effects, through which the researchers can determine the specific mediating effect for each mediator (Preacher & Hayes, 2008). Each test in this study used a α level of ≤.05 to determine statistical significance.

Results

Table 2 presents the correlation results among studied variables. Being female, younger age, lower education, poorer self-rated health, lower self-rated class and lower job satisfaction were related to more depressive symptoms. Each of the four community factors was significantly related to depressive symptoms in the expected direction. Higher levels of community cohesion, lower community threat, larger community supportive network size and higher levels of medical insurance coverage were related to lower depressive symptoms.

Table 2.

Bivariate correlates of depressive symptoms.

	2	3	4	5	6	7	8	9	10	Depressive symptoms
1. Age	−.027**	−.386**	−.332**	−.028**	.038**	.203**	−.092**	−.046**	−.003	−.019**
2. Female		−.085**	−.060**	.021**	−.008	−.038**	−.021**	−.101**	−.021**	.061**
3. Education			.239**	.135**	.103**	−.247**	.115**	.155**	.461**	−.052**
4. Self-rated health				.219**	.099**	.038**	−.063**	.142**	.052**	−.225**
5. Self-rated class					.202**	.078**	−.116**	.144**	.087**	−.151**
6. Self-perceived job satisfaction						.087**	−.083**	.076**	.089**	−.235**
7. Community cohesion							−.205**	.209**	−.223**	−.109**
8. Community threat								−.020**	.101**	.122**
9. Supportive network size									.109**	−.061**
10. Medical insurance coverage										−.064**

Significance at .05 level.

Significance at the .01 level (two-tailed).

Table 3 shows the results of two regression models. In the first model, only control variables and rural versus urban were included. Rural and urban differences remained significant, with rural workers having higher scores of depressive symptoms (B = 0.149, p < .05). In the second model, four mediators were added and indirect effects of residential status were estimated through Hayes’ SPSS Process macros. Results indicate that rural and urban differences in depressive symptoms were no longer significant (B = 0.091, p > .05). Each of the mediators except community supportive network size had significant direct effects on depressive symptoms. The indirect effect of residential status via community cohesion was −.115, with 95% confidence interval [−0.159, −0.072], via community foreseeable threat (B = −0.077) with 95% confidence [−0.099, −0.058] and via medical benefit coverage (B = 0.249), with 95% confidence [0.170, 0.320]. The negative indirect effects through community cohesion and community threat indicated that rural–urban differences were lessened because rural communities were better than urban communities in terms of community cohesion and community threat. The positive indirect effects through medical benefit coverage indicated that this community factor had widened rural–urban gaps in depressive symptoms because rural workers had inferior medical insurance coverage compared to their urban counterparts.

Table 3.

Direct and indirect effects of residential status on depressive symptoms.

	Model 1		Model 2
	B	SE	B	SE
Rural	0.149*	0.068	0.091	0.082
Controls
Age	−0.023**	0.002	−0.019**	0.002
Female	0.402**	0.053	0.414**	0.053
Education	0.018	0.015	0.029	0.016
Self-rated health	−0.816**	0.029	−0.779**	0.029
Self-rated class	−0.164**	0.016	−0.144**	0.017
Self-perceived job satisfaction	−0.871**	0.033	−0.828**	0.034
Mediators
Community cohesion			−0.061**	0.011
Community foreseeable threat			0.069**	0.008
Supportive network size			0.006	0.012
Medical benefit coverage			−0.160**	0.027
Indirect effects
Via community cohesion			−0.115**	0.022
Via community foreseeable threat			−0.077**	0.011
Via supportive network size			0.001	0.001
Via medical benefit coverage			.249**	0.041
Adjusted R²	0.11			0.12

SE: standard error.

The change in adjusted R² values form Model 1 to Model 2 was significant (p < .01).*Significance at .05 level.

Significance at .01 level.

To better illustrate the indirect effects, we plotted the coefficients of the paths among rural–urban residency, community mediators and depressive symptoms (see Figure 2).

Figure 2.

Results of estimates of direct effects of urban–rural and mediators in a path diagram.

Discussion

Our findings indicated that rural workers have more depressive symptoms than their urban counterparts do and identified community factors that contributed to the rural–urban differences in depressive symptoms. Among the four mediators, community cohesion, foreseeable community threat and medical benefit coverage each mediated the difference between the rural and the urban workforce. Specifically, community cohesion and foreseeable community threat lessened rural–urban differences; medical benefit coverage widened such differences. This suggests that the mechanisms that may account differences in depressive symptoms between urban and rural workers are multifaceted and may run in different directions. Our findings further substantiate that depressive symptoms are not only the reaction of personal inner emotions but also may be affected significantly by the external environment. Furthermore, the environmental factors identified appear somewhat malleable to interventions. Improving the mental health of the labor force in China therefore requires attention not only to the personal conditions of workers but necessitates strategies to create a community environment of trust, security and equality.

Unlike the urban community, China’s rural communities are built on kinship and geographic relationships – a typical society of acquaintances. Rural areas not only inherited many traditions of mutual help and cooperation (Hao, 2010) but also can increase the individual’s sense of security and mutual trust due to low mobility (Li & Zhong, 2013). Our study found that community cohesion captures the relational features of a community, and in this regard, rural communities are better off than urban communities. This rural advantage appears to reduce the rural–urban disparities in depressive symptoms. This finding can inform policymakers about the mental health consequences resulting from loss of community cohesion as urbanization in China has advanced. One implication is the possibility of facilitating the establishment of a ‘quasi-acquaintance society’ that promotes positive and trustful neighborhood relationships in urban communities. The ‘quasi-acquaintance community’ model tested in Shenzhen, a southern city in China, has shown positive effects in increasing community unity and promoting mutual trust in residents (Li, 2015).

Community threat measures persons’ assessment of risky and potentially harmful features. Although there is a gap between real risk and personal perceived risk, the impact of risk perceptions on the public psychology is ubiquitous. High-risk perception might cause psychological panic, helplessness and anxiety (Li, Fan Mejia, Wang, & Hao, 2009). In China, farmers are considered to occupy a profession that never suffers from unemployment. Compared to people in urban areas, rural populations have fewer concerns regarding the risk of the natural environment and food safety, due to low population density and low risk of infectious diseases. Thus, there are significant differences in foreseeable community threats between urban and rural areas. Low-risk perception among the rural workforce in turn can reduce the differences in depressive symptoms between rural and urban labors. It may be possible to address depressive issues in the rural labor force by building in these positive rural community strengths, such as by maintaining the mutual trust between rural community residents, improving the sanitation conditions of rural communities and strengthening the support for rural economic construction to stabilize the income of rural labor.

The third significant mediator affecting the differences in depressive symptoms between rural and urban workers is medical benefit coverage. It had the highest mediating effect size among the three significant mediators. Previous studies have shown that a sound healthcare system can effectively promote personal physical and mental health (Tang et al., 2014). Since the collapse of the old cooperative medical system in 1978, most of China’s rural workforce did not receive effective medical insurance for a long period. In 2003, the new rural cooperative medical demonstration projects started to be pilot tested in some rural counties, and coverage eventually expanded across all rural areas in China in 2010. Despite its availability in all rural areas, the covered benefits of the new rural cooperative system remained lower and less adequate than the three tier healthcare coverage available in urban areas. The three tier healthcare system in urban areas consists of public health care, urban worker basic medical system and urban resident basic medical insurance system, which all have better coverage than the new rural cooperative system (Wang, Su, Yan, & Zhang, 2009). The new rural cooperative medical system needs to be improved in many aspects, such as hospital and clinic accessibility, covered diseases, drug benefits, copayments and deductibles. To reduce the rural–urban disparities in depressive symptoms, the government not only needs to improve the current rural cooperative medical system for the rural workforce but also aim to eradicate the structural segmentation rooted in the current household registration system that resulted in the fragmentation of the medical insurance system.

The community supportive network size did not have a significant effect on depressive symptoms in the multivariate model. However, it is worthy of some discussion. China is a highly relationship oriented society, in which both rural and urban residents form a certain social network based on communities. These social networks are not only an important platform for information sharing, but they also provide alternative mechanisms for resource allocation (Li & Chen, 2012). In addition, having a ‘network of friends’, a ‘good social life’ and ‘being liked’ particularly demonstrates the interconnectedness of social factors linked to ‘self confidence’ (Crossley & Langdridge, 2005), which affects the mental health of residents. We found that the rural workforce had smaller community supportive network than their urban counterparts. This may be due to the relative incapability of rural workforce to create, maintain and use community relation networks than the urban workforce. It alternatively could be due to the impact of cultural change, including the weakening of original rural social networks that enervate the already fragile social relation networks built on geographical and kinship (Mao, 2013). Moreover, we are limited to measuring only networks that can offer tangible help (e.g. financial assistance). It is possible that social network patterns would be different if emotional or intangible support or help could be measured as well, which is an area of interest for future research. One implication we can draw is that policymakers should not assume that rural workers living in communities ascribed to traditions and cultural values will automatically have large supportive networks. Instead, deeper investigations are needed to understand the type, level and quality of the support rural workers can obtain, and solutions or strategies need to be based upon findings of these investigations before any interventions or services aimed to promote community support can be recommended.

Several limitations in this study should be noted. First, the four community features covered essential community structure and relational characteristics, but we were limited to secondary data available and perhaps excluded other important community factors, such as community heterogeneity and community economic development, which can be explored in future studies. Second, community variables were assessed through individual’s perceptions, and measures assessing community aggregate features should be added in the future. Finally, as a cross-sectional study, it cannot rule out the possible influence of individual mental health on perceptions of the social factors (especially the subjective assessment of their own community factors). That is, it is possible that workers with poor mental health may perceive community social relations differently because of their mental health status, and the perceived risks of or isolation from the community. This possibility could be better examined with the availability of more objective community measures.

In conclusion, this study reveals that having lower education, poorer self-rated health, lower self-rated class, lower job satisfaction and being female and younger age all were related to more depressive symptoms. As we assumed, higher levels of community cohesion, lower community threat and higher levels of medical insurance coverage were related to lower depressive symptoms. Community factors not only influenced depressive symptoms directly but also mediated the residential differences on depressive symptoms. This study sheds light on distinctive roles of community factors in explaining rural–urban differences in depressive symptoms. Rural workers are better off living in a community with higher levels of community cohesion and less foreseeable threat, but subject to smaller support network size and limited medical coverage. Policies or programs should be designed to promote rural community strengths and to address their weaknesses.

Footnotes

Funding

The author(s) received no financial support for the research, authorship and/or publication of this article.

References

Berkman

L. F.

Syme

S. L.

(1979). Social networks, host resistance, and mortality: A nine-year follow-up study of Alameda County residents. American Journal of Epidemiology, 109, 186–204.

Chen

S. L.

X. L.

(2006). The rural citizen’s psychological status and investigation approach. Chinese Medical Ethics, 5, 117–118.

Chi

(2014). Social networks, psychological capital and health comparison of urban and rural residents. Population & Development Review, 20(3), 96–103.

Cole

M. G.

Dendukuri

(2003). Risk factors for depression among elderly community subjects: A systematic review and meta-analysis. American Journal of Psychiatry, 160, 1147–1156.

Crossley

Langdridge

(2005). Perceived sources of happiness: A network analysis. Journal of Happiness Studies, 6, 107–135.

Cui

Wang

Liu

Jing

Liu

Zou

(2014). Relationship between occupational stress and depression in migrant workers. Chinese Journal of Industrial Hygiene and Occupational Diseases, 32, 112–114.

Ding

(2009). Evaluating the mental health status of rural residents based on Scale of Kessler 10. Chinese Health Service Management, 9, 23.

Fiori

K. L.

Antonucci

T. C.

Cortina

K. S.

(2006). Social network typologies and mental health among older adults. The Journals of Gerontology Series B: Psychological Sciences and Social Sciences, 61(1), 25–32.

Gao

X. M.

J. H.

(. (2015). Causality research on socio-economic status and health inequality. Journal of Shenzhen University (Humanities & Social Sciences), 32(6), 45–53.

10.

Hao

Y. G.

(2010). From the village to the community: The logic of the transformation of government functions – From the perspective of production socialization. Socialism Studies, 190(2), 68–72.

11.

X. S.

Keung

W. F.

Shouchui

(2010). Rural-urban migration and mental health: Evidence from Shanghai. Sociological Studies, 1, 111–129.

12.

Y. H.

(2015, November 15). Old Guangdong do not believe ‘no reason is Germany’ and hold the belief of higher education more happiness about women. Yancheng Evening News, p. A14.

13.

Kennelly

O’Shea

Garvey

(2003). Social capital, life expectancy and mortality: A cross-national examination. Social Science & Medicine, 56, 2367–2377.

14.

Lewin

(1951). Field theory in social science: Selected theoretical papers (D Cartwright, Ed.). New York, NY: Harper.

15.

Zhong

Z. B.

(2013). A Study on the transformation of the mature society in the background of farmers’ concentrated living. Academic Journal of Zhongzhou, 5, 74–78.

16.

H. Q.

Fan Mejia

J. M.

Wang

S. H.

Hao

L. G.

(2009). The Public perception of risks and the management of emergency measures taken during unexpected calamities [J]. Management World, 6, 8.

17.

J. W.

(2015). The practical study of social work’s interference with the construction of ‘Acquaintances Community’ – Taking community F in Shenzhen City as an example (Unpublished master’s thesis). Zhengzhou University, Zhengzhou, China.

18.

Chen

(2012). Can ‘relationship’ bring happiness? Empirical evidence from rural China. China’s Rural Economy, 8, 66–78.

19.

Liu

(2016). China’s income distribution system reform and income growth strategy. Open Access Library Journal, 3(05), 1.

20.

Lowry

Xie

(2009). Socioeconomic status and health differentials in China: Convergence or divergence at older ages? (Population Studies Center). Ann Arbor: University of Michigan. Retrieved from https://scholar.princeton.edu/sites/default/files/yuxie/files/lowry-xie2009.pdf

21.

Mao

D. S.

(2013). On the formation of social relation network and the difference of happiness of urban and rural residents. Business Economics Research, 29, 16–18.

22.

Preacher

K. J.

Hayes

A. F.

(2008). Asymptotic and resampling strategies for assessing and comparing indirect effects in multiple mediator models. Behavior research methods, 40(3), 879–891.

23.

J. B.

Wang

Meng

Q. Y.

(2008). Relationship between social capital and health outcome in rural China. Chinese Primary Health Care, 1, 11.

24.

Snelgrove

J. W.

Pikhart

Stafford

(2009). A multilevel analysis of social capital and self-rated health: Evidence from the British household panel survey. Social Science & Medicine, 68, 1993–2001.

25.

Subramanian

S. V.

Kim

D. J.

Kawachi

(2002). Social trust and self-rated health in US communities: A multilevel analysis. Journal of Urban Health, 79(4 Suppl. 1), S21–S34.

26.

Tang

J. L.

Wang

J. L.

W. Y.

(2014). Mental health of elderly people in different medical security system. Chinese Rural Health Service Administration, 34(1), 63–67.

27.

Wang

J. Y.

Yan

L. L.

Zhang

(2009). Comparison of medical insurance system of urban residents and new rural cooperative medical system based on line fair perspective. Medicine and Society, 22(1), 3–5.

28.

Wang

(2012). Historical development and future prospects of household registration reform in the context of urbanization. Urban Management of Shanghai, 2, 41–45.

29.

Wong

D. F. K.

Leung

Lau

Chang

(2008). Mental health of migrant workers in China: Prevalence and correlates. Social Psychiatry and Psychiatric Epidemiology, 43, 483–489.

30.

Xue

Y. F.

Wang

X. L.

(2005). An investigation on the residents’ physical and psychological health states in Hebei province. Chinese General Practice, 21, 1762–1764.

31.

Zhao

(2008). Social network and people’s wellbeing in urban and rural areas. Society, 5, 1–19.