One Religion, Two Tales: Religion and Happiness in Urban and Rural Areas of China

Abstract

Most previous studies performed in Western social contexts have revealed that religion can influence an individual’s sense of happiness. Few studies have sought to clarify the influence of religion in a Chinese social context, however, and there has been no study specifically about the potential differences in the dichotomous social environments of urban and rural areas in China. Via the nationwide survey data of the 2007 Spiritual Life Study of Chinese Residents (SLSC), this study examines the association between religion and happiness among urban and rural residents of China. The results reveal that there is a generally positive association between religion and happiness among those with religious affiliations in China. Regardless of affiliation with Buddhism or Protestantism, there is a strong positive association among rural respondents, an insignificant association among urban respondents, and mixed results among town residents.

1. Introduction

Happiness, a feeling that is desired by every human being (Rizvi and Hossain 2017), is a “psychological outcome that encompasses an individual’s cognitive judgments of satisfaction and the affective appraisals of emotions” (Kesebir and Diener 2008, p. 118). A lot of studies have determined a long list of factors relating to the how and why of happiness. Factors such as health, marital status, income, social status, psychological status, social networks, and social capital can influence the happiness of an individual (Chen and Williams 2016; Cheung 2016). Among these factors, religion has been one of the unique factors found to produce the benefits of mental health, physical well-being, social support, and internal peace that ultimately lead to an individual’s happiness (Rizvi and Hossain 2017).

Meanwhile, scholars have proposed that among different cultural contexts, such as those found in East Asia, South Asia, and the Middle East (e.g., Lim et al. 2019; Devine et al. 2019; Fariddanesh and Rezaei 2019), there has only been a very small volume of literature focusing on countries where atheism is the social norm (McCullough and Willoughby 2009; Potter 2003; Wielander 2017; Thege et al. 2013; Yang 2006). China is considered to be much less religious than many Western countries (Overmyer and Edmonds 2003; Wang et al. 2015). However, China has a long history of religion, and there has been a religious awakening in China during the past decades (Chau 2010; Liu et al. 2011; Wei and Liu 2013; Ying et al. 2017). Existing studies on the relationship between religion and happiness in mainland China have presented some mixed results (Deng et al. 2019; Liu and Koenig 2013; Chen and Williams 2016; He et al. 2016; Lu and Zhang 2016; Zhang et al. 2016).

Further, the secularization hypothesis of religion (Norris and Inglehart 2011) argues that conditions of socioeconomic inequality may influence religiosity, and that social environment may strengthen or attenuate associations between religion and happiness (Büssing and Koenig 2010; Diener et al. 2011; Lun and Bond 2013). China has long been a place of social dichotomies, with significant disparities between social economic development levels and individual well-being between urban and rural residents (Knight and Gunatilaka 2010; Knight et al. 2009; Liang and Wang 2014; Treiman 2012; Zhao 2012). Constituting more than 50% of the whole population, rural residents (peasants) are usually considered to be of lower social status in China due to their much lower income level and poor living conditions. Urban and rural residents may also have different individual or social motivations to practice religion, and despite the hostile rural social environment, the religious residents in China might be more likely to report higher levels of happiness than their peers living in urban areas. Using the nationwide survey data of the 2007 Spiritual Life Study of Chinese Residents (SLSC), this study examines the association between religious involvement and self-reported happiness among the sampled residents of China, and further examines whether the religious urban and rural residents are equally likely to report being happy, controlling for their demographic and socioeconomic characteristics.

2. Religion and Happiness

2.1. The Dichotomy of China: Urban and Rural Areas

Socioeconomic environment. China is a dichotomous society and this societal issue is not only demonstrated in the high level of unequal mean income between rural residents and urban residents, but also in the serious lack of services such as education, social security benefits, and medical care in the rural areas. Despite the great improvements of living standards relative to decades ago, many rural areas in China are still lacking health care services, social security benefits, pension funds, and social capital (Chen and Williams 2016).

Social values, life style and social status. Urban residents emphasize an urbanized life style, equal social status, and greater liberty. Conversely, rural residents emphasize traditional life values and life styles (Davey et al. 2009). Compared to urban residents, the social status of rural residents (peasants) is considered to be at the lower level(s) in the social hierarchy of China (Davey et al. 2009).

Gap of happiness. The socioeconomic dichotomy of urban–rural areas has generated a large gap of happiness among residents, too (Knight et al. 2009; Liang and Wang 2014). On the one hand, in urban areas, “progress paradox” documents the phenomenon that people within an environment of rapid economic growth encounter insecurity (life environment changes) and inequality (winners and losers in the changing process) (Graham and Pettinato 2002). As per these results, urban educated residents in China, compared with rural residents, are more likely to report life stresses, depression, poor mental health, or lower life-satisfaction (Graham et al. 2017).

Based on the unique characteristics of the Chinese social context, it is reasonable to predict that low income and life quality will result in rural residents having comparatively lower levels of happiness than urban residents. However, some scholars argue that as long as a rural resident’s basic needs are satisfied and they prefer a materialistically simpler life, then their happiness levels will not be significantly different from those of urban residents (Davey et al. 2009; Diener and Biswas-Diener 2002; Biswas-Diener et al. 2005).

2.2. Motivations for Religious Involvement in Urban and Rural China

The ratio of the population who are regularly engaged in religious activities has increased from 7% in 2001 (Stark and Liu 2011), to 10% in 2012, and to 26.4% in 2014 (Lu and Zhang 2016; Wenzel-Teuber 2017). Within the dichotomy of the social environment in China, residents in urban and rural areas of China have heterogeneous motivations for participation in religious activities.

In urban areas of China, the motivation for religious conversion may largely be due to the psychological response to social anomies, inequality (Chen and Williams 2016; Lu 2012), or living stresses (Graham et al. 2017). By emphasizing non-material aspects of life, religion may help to improve happiness by providing stress coping skills, as well as by providing a strong sense of purpose and acceptance over and above one’s economic conditions or social status (Chen and Williams 2016).

In rural China, the major reason for religious conversion apart from the religious tradition may be due to family members becoming seriously sick (Liang and Qi 2015; Zheng and Wang 2014) or suffering from some chronic disease (Zhou and Yu 2017). Given the limited reimbursement under the rural health insurance system in China, chronic diseases result in big economic and time costs to rural families. Those facing the hardship of a health crisis may turn to religion, especially Christian involvement in rural China, as an alternative healing method, as they search for social or psychological supportive resources (Zhou and Yu 2017).

2.3. Association and Mechanism of Religion and Happiness

Happiness refers to “a person’s cognitive and affective evaluations of his or her life” (Diener et al. 2002, p. 63). The benchmark of happiness is rooted in the mind of the individual and determines whether he or she is happy (Lim et al. 2019; Lu et al. 2001). Researchers argue that both objective (e.g., income) and subjective factors (e.g., values and religion) can influence the happiness of an individual. Furthermore, whether one objective factor can influence an individual’s happiness depends on whether or not it can change the individual’s subjective state of mind in response to it (Lim et al. 2019; Reyes-García et al. 2016).

Religion is a subjective belief (or set of beliefs) that influences the desired life values, life styles, and life purposes of religious people (Cohen and Johnson 2017; Devine et al. 2019; Fariddanesh and Rezaei 2019; Lun and Bond 2013). Thus, religion influences religious peoples’ subjective responses to objective factors.

The previous study achieved fruitful results suggesting that religion has a positive effect on happiness for people who are religious, providing better physical and mental health and thus greater happiness (Büssing and Koenig 2010; Cohen and Johnson 2017; Krause and Wulff 2005; Lewis and Cruise 2006; Sander 2017; Rizvi and Hossain 2017). For instance, the intrapersonal aspects of religion, such as prayer, meditation, and self-reflection, associate positively with happiness (Greeley and Hout 2006; Krause 2003). Furthermore, religious faith enhances the happiness of individuals via the provision of a comprehensive framework for the interpretation of world events and life challenges, and thus a sense of meaning and purpose in life (Inglehart 2010; Lim and Putnam 2010).

Religion may impact an individual’s happiness through various mechanisms such as social capital (Chen and Williams 2016), social networks (Lim and Putnam 2010), lifestyle or health-related behaviors (Diener and Seligman 2002), and positive psychological status (e.g., faith, inner peace) (Myers 2000; Stark and Maier 2008). Some studies also reveal that being a member of a religious community helps an individual develop longstanding close relationships with other members. These relationships will lead the individual to experience a sense of attachment, commitment, social support, and happiness (Krause 2003). Individuals who turn to religion will obtain social support (e.g., friendship and community), psychological resources (e.g., self-identification, self-esteem, and self-efficacy), coping strategies for dealing with stress or trauma (e.g., illness, unemployment, divorce), a sense of meaning and purpose for life in an unpredictable world, and an expected utility in the afterlife (Brown and Tierney 2009; Greeley and Hout 2006; Inglehart 2010; Krause 2003; Lim and Putnam 2010).

2.4. Association of Religion and Happiness in Urban and Rural China

The secularization hypothesis of religion (Norris and Inglehart 2011) argues that patterns of human security and, in particular, conditions of socioeconomic inequality drive religiosity. People are religious in countries where there are low quality and/or ineffective public services, incomes are low, and personal insecurity is high. In these hostile environments, “religion provides a partial refuge, from disease and distress when it can, and a promise of relief in the next world when it cannot” (Deaton and Stone 2013). Further, when the socioeconomic environment has improved and the citizens can obtain better economic and personal security, religiosity often declines (Norris and Inglehart 2011).

Based on the secularization hypothesis of religion, we can infer that rural religious residents in China probably report higher levels of happiness than those of the urban residents, because religion can better help them in the hostile rural social environment of China. Meanwhile, religiosity among urban residents is not as strong as predicted by the secularization hypothesis (Norris and Inglehart 2011).

In summary, we propose the hypothesis of this study as follows:

Hypothesis 1. 

Religious affiliation is generally positively associated with the self-reported happiness of Chinese urban and rural residents.

Hypothesis 2. 

The association between religion and happiness among rural residents may be stronger than those of urban residents.

3. Data and Sampling

Data comes from the 2007 Spiritual Life Study of Chinese Residents (SLSC), a national multi-stage probability sample of 7021 Chinese in mainland China. A high-quality team of international scholars, with Dr. Fenggang Yang as the principal investigator, designed the questionnaire and led the project. HorizonKey, a Chinese research firm, collected the data through face-to-face interview surveys.

Until now, the SLSC dataset has been the only nationally representative survey with rich information regarding the religious and spiritual lives of residents in mainland China. No new comprehensive survey about religious life in China has been conducted since the SLSC in 2007, and no new survey is planned for the near future. Despite the rapid economic and social changes during the past decade, there have been no remarkable changes with respect to the roles of religion in China since 2007 (Lu and Gao 2017). Therefore, the SLSC dataset should still provide relevant information for understanding religious life in contemporary China (Lu and Gao 2017). The Chinese General Social Surveys (CGSSs) provide more updated samples of 2012 to 2015; however, they only contain three general questions about the religion of a respondent and the frequency of religious attendance.

Respondents had to be 16 years of age or older, have lived at their current residence for three months, and not have done a survey within the past six months. Also excluded were people with jobs related to surveys, marketing investigation, public relations, or the media. Respondents were selected using a multi-stage method from select metropolitan cities, towns, and administrative villages. The final survey was administered in 56 locales throughout China, including three municipal cities (Beijing, Shanghai, Chongqing), six provincial capital cities (Guangzhou, Nanjing, Wuhan, Hefei, Xi’an, and Chengdu) 11 regional-level cities, 16 small towns, and 20 administrative villages. Within each locale, households were sampled within neighborhoods, and neighborhoods were sampled within administratively defined total neighborhood committees (government-defined collections of neighborhoods). A KISH grid procedure was used to randomly select one respondent from each household for a face-to-face in-home interview. The dataset contained sampling weights to reflect the general population parameters in the 2006 Statistical Yearbook of China.

Starting with 7021 observations in the original dataset, we dropped the observations from respondents who “had difficulty in understanding the survey questions”, respondents who refused to answer or felt it hard to speak about their religious beliefs, as well as observations with missing values for key variables. The final dataset used in this study contained 6267 valid responses, accounting for 89.26% of the original dataset.

4. Statistical Analysis Method and Variables

4.1. Dependent Variable: Self-Reported Happiness (SRH)

The dependent variable in this study was self-reported happiness. The survey measured the subjective happiness by asking respondents “Do you feel happy about your life overall? Are you very happy, somewhat happy, somewhat unhappy, or very unhappy?” Responses to this questions were ranked on a five-point scale from 1 (very unhappy) to 5 (very happy). This global measure of happiness as subjective well-being has been shown to be valid in large sample surveys (Brockmann et al. 2009; Layard 2011; Reyes-García et al. 2016). The same measure of subjective well-being has also been adopted by the CGSS database of China and has yielded broadly consistent results in the literature (Chen and Williams 2016).

4.2. Explanatory Variables

Religious affiliation. The religious affiliation of the respondents was identified by a question “Regardless of whether you have been to churches or temples, do you believe in a religion?” More than 20% of the respondents reported a religion they believed in. If the answer was “Yes”, the respondent was further asked to indicate the specific affiliation. Buddhism, Protestantism, Islam, Catholicism, Daoism, and Confucianism were recorded in the original dataset of the SLSC. As expected, Buddhism was the largest group, and accounted for about 17% of the final sample population. Protestantism was the second largest, and accounted for about 2.41%. Together, Buddhists and Protestants accounted for about 94.4% of individuals with religion affiliations. Each of the other religious affiliations (Islam, Daoism, Catholicism, and Confucianism) accounted for less than 0.5% of the sample population. Because the observation numbers of these religious affiliations were insufficient for conducting valid regression analyses (Chen and Williams 2016), we did not identify religious affiliations other than Buddhism and Protestantism. Two dummy variables indicated a respondent claiming to be a believer of “Buddhism” or “Protestantism”.

Importance of religion. An interview question asked respondents “Please tell me the importance of the following items in your life. Is it very important, somewhat important, somewhat unimportant, or not at all important?” Religion is listed together with other items such as family, friends, entertainment, politics, and career. The importance of religion is recorded with a four-point scale from 1 (very important) to 4 (not important at all). To capture the potential effect clearly, we constructed a dummy variable to identify respondents who considered religion “very important”.

Frequent religious attendance. An interview question first asked respondents “Did you worship in a conventional religious setting (regardless of it being a temple or a church building) during past 12 months?”, then asked about the frequency as “regularly or occasionally”. We only identified respondents who had regular religious attendance. The base group is “have no or occasional religious attendance”.

Ever practice praying. An interview question asked “Do you ever pray? Have you ever tried to communicate with God or a certain supernatural power, asking for blessings and protection?” A dummy variable took the value of 1 if the answer was “yes”.

Residency. This survey dataset had a variable indicating if the respondents lived in an urban, town, or rural setting at the time of the survey. A qualified respondent should have continuously occupied the residency during the past three months. According to the Rural Urban Classification of China (National Bureau of Statistics of China 2010)1, urban areas (code 100) include cities, counties and townships established according to the Administrative Division System of China. Town (code 210) is a central area in rural environment, serving as a regional economic, cultural, and service center without an administrative establishment. Rural (code 220) refers to the villages and rest areas in the non-urban environment. Since town areas are the transition areas between the typical urban and rural environments, we adopted a dummy variable of “town area” to control the potential mixed effects in these areas, so that the net effects in typical urban and rural areas could be clearly captured.

4.3. Control Variables

Socio-demographic variables were included to control for their potential confounding effects on the self-reported happiness. The survey dataset contained information about respondents’ genders, ages (measured in years), ethnic groups, marital statuses, education levels, occupations (unemployed status included), household economic statuses, and self-rated health statuses.

This study adopted self-rated health as a measure of general health status. A respondent’s self-rated health status is widely considered as a valid proxy for health and a reliable predictor of future health outcomes such as morbidity and mortality (Idler and Benyamini 1997; Stavrova 2015).

To control for the impacts of social capital on the well-being of respondents, we constructed a dummy variable of a social capital indicator based on the survey question “Did you attend any of the following events in the past twelve months?” The events are listed as attending other people’s weddings, charity events, various social events, vacations, or entertainments events with family or friends and so on. If the respondent said “yes” to any of the events listed above, the social capital indicator variable takes the value of 1.

4.4. Statistical Analysis Method

Ordered logistic regression may be applied when a dependent variable has more than two categories and the values of each category have a meaningful sequential order where a value is indeed “higher” than the previous one. In this study, the dependent variable was the five-point variable “happiness”. We performed model tests in STATA, and an insignificant test statistic indicated that the final model did not violate the proportional odds/parallel lines assumption. Therefore, we could apply the command svy: ologit in STATA 14 statistical package (Stata Corp LP, College Station, TX, USA) to analyze the final sample.

First, the baseline model of our empirical analysis examined the relationship between happiness and general religious beliefs, controlling for potential demographic and socioeconomic confounding factors. We also examined how self-reported happiness varied among different religiously affiliated groups.

Second, to examine how an association with happiness may vary among respondents in urban and rural areas, we included the interaction terms of religious affiliation and residential status (urban and rural).

As a robustness check, we replaced religious affiliations with variables measuring inner religiosity (“importance of religion”) and religious practice (“frequency of religious attendance”, “practice praying”). The interaction terms of these religiosity measurements and residential statuses (urban and rural) were also constructed, respectively.

5. Results and Analysis

Table 1. Descriptive Statistics.

Variables		Full Sample		Urban		Town		Rural
		N	%	N	%	N	%	N	%
Happiness	Very unhappy	24	0.38	13	0.42	3	0.18	8	0.52
	Unhappy	198	3.16	89	2.89	37	2.24	72	4.69
	So-so	961	15.33	496	16.10	218	13.20	247	16.09
	Happy	2980	47.55	1583	51.40	727	44.01	670	43.65
	Very happy	2104	33.57	899	29.19	667	40.38	538	35.05
Self-reported Health	Very unhealthy	33	0.53	15	0.49	5	0.30	13	0.85
	Unhealthy	238	3.80	96	3.12	46	2.78	96	6.25
	So-so	632	10.08	353	11.46	133	8.05	146	9.51
	Healthy	2109	33.65	1176	38.18	466	28.21	467	30.42
	Very healthy	3255	51.94	1440	46.75	1002	60.65	813	52.96
With religion	Yes	1291	20.62	612	19.87	358	21.67	321	20.91
Religious	Buddhism	1068	17.04	500	16.23	299	18.10	269	17.52
affiliation	Protestantism	151	2.41	73	2.37	34	2.06	44	2.87
	Islam	30	0.48	13	0.42	17	1.03	0	0.00
	Daoism	19	0.30	11	0.36	3	0.18	5	0.33
	Catholicism	12	0.19	8	0.26	2	0.12	2	0.13
	Confucianism	11	0.18	7	0.23	3	0.18	1	0.07
Importance of religion	Very important	175	2.79	68	2.21	62	3.75	45	2.93
	Somewhat important	586	9.35	305	9.90	159	9.62	122	7.95
	Somewhat unimportant	1296	20.68	692	22.47	347	21	257	16.74
	Not important at all/Don’t know	4210	67.18	2015	65.42	1084	79.00	1111	72.38
Have you ever prayed?	Yes	710	11.33	404	13.12	185	11.20	121	7.88
Worship regularly	Yes	185	2.95	79	2.56	75	4.54	31	2.02
Demographic and socioeconomic characteristics
Gender	Male	3005	47.95	1479	48.02	800	48.43	726	47.30
Age *		39.87	(13.60)	40.12	(13.74)	38.10	(12.78)	41.28	(14.05)
Marriage status	Married/cohabiting	5085	81.14	2434	79.03	1339	81.05	1312	85.47
	Divorced/separated/widowed	264	4.21	147	4.77	60	3.63	57	3.71
	Unmarried	918	14.65	499	16.20	253	15.31	166	10.81
Ethnic group	Han (Yes)	6009	95.88
Education level	No schooling	236	3.77	72	2.34	36	2.18	128	8.34
	Elementary	717	11.44	174	5.65	143	8.66	400	26.06
	Junior middle school	2082	33.22	839	27.24	583	35.29	660	43
	High school	2074	33.09	1200	38.96	583	35.29	291	18.96
	College and above	1158	18.48	795	25.81	307	18.58	56	3.65
Economic Status	Lower	728	11.62	353	11.46	145	8.78	230	14.98
	Middle	3944	62.93	1987	64.51	1019	61.68	938	61.11
	Middle-high	1477	23.57	695	22.56	448	27.12	334	21.76
	High	118	1.88	45	1.46	40	2.42	33	2.15
Unemployed	Yes	344	5.49	221	7.18	123	7.45	0	0
Social capital	Yes	4413	70.42	964	31.30	400	24.21	490	31.92
Regions in China	East China	319	5.09	152	4.94	98	5.93	69	4.50
	South China	1011	16.13	153	4.97	321	19.43	537	34.98
	Central China	1246	19.88	553	17.95	318	19.25	375	24.43
	North China	996	15.89	725	23.54	191	11.56	80	5.21
	Northwest China	268	4.28	160	5.19	108	6.54	304	19.80
	Southwest China	1020	16.28	431	13.99	285	17.25	170	11.07
	Northeast China	897	14.31	396	12.86	331	20.04	-	-
	Beijing Area	257	4.10	257	8.34	-	-	-	-
	Shanghai Area	253	4.04	253	8.21	-	-	-	-
Total		6267	100	3080	100	1652	100	1535	100

*: Age is a continuous variable with the means reported and the standard deviation in parentheses.

reports the descriptive statistics of the dataset. The subsamples of with/without religions are divided according to the religious affiliations reported. Across the various demographic and socioeconomic variables, there was no significant difference between the groups with religion and without religion. This is despite the fact that the religious group was a population with a 3% higher elementary education level and contained 3.78% fewer respondents with a college-and-above education. The results also show that the religious group was more likely to have social capital. Respondents in East China, South China, or Shanghai City were more likely to report themselves to have religious beliefs than those from the rest of the areas in China. This difference was due in part to the historical fact that Protestantism originally spread into China from eastern and southern coastal cities.

Table 2. Association between happiness and religion (basic model; ordered logistic model, odds ratio reported).

	(1)	(2)	(3)
Variables	Happiness	Happiness	Happiness
Ref. Group	Urban residents	Urban residents w/t religion	Urban residents w/t religion + other affiliations
With Religious affiliation		1.227 *** (0.0870)
Buddhism			1.175 ** (0.0824)
Protestantism			1.399 ** (0.231)
Town resident	0.990	0.982	0.991
	(0.164)	(0.163)	(0.164)
Rural resident	0.718 **	0.714 **	0.715 **
	(0.0998)	(0.0993)	(0.0994)
Health	1.972 ***	1.967 ***	1.970 ***
	(0.0649)	(0.0648)	(0.0649)
Social Capital	1.117 *	1.103 *	1.102 *
	(0.0648)	(0.0642)	(0.0643)
Middle economic status	1.707 ***	1.698 ***	1.700 ***
	(0.141)	(0.140)	(0.140)
Middle-high economic status	2.654 ***	2.639 ***	2.644 ***
	(0.257)	(0.255)	(0.256)
High economic status	3.747 ***	3.731 ***	3.760 ***
	(0.797)	(0.794)	(0.800)
Unemployed	0.905	0.906	0.906
	(0.102)	(0.102)	(0.102)
Age	1.002	1.002	1.002
	(0.00251)	(0.00251)	(0.00251)
Male	0.792 ***	0.800 ***	0.802 ***
	(0.0404)	(0.0409)	(0.0410)
Divorced, separated, or widowed	1.146	1.148	1.142
	(0.0966)	(0.0967)	(0.0962)
Unmarried	0.514 ***	0.511 ***	0.510 ***
	(0.0785)	(0.0780)	(0.0778)
Han Ethnic Group	1.209	1.231	1.203
	(0.176)	(0.179)	(0.175)
Elementary	0.736 **	0.734 **	0.736 **
	(0.112)	(0.112)	(0.112)
Junior middle school	0.802	0.806	0.809
	(0.117)	(0.118)	(0.118)
High school	0.945	0.952	0.957
	(0.144)	(0.145)	(0.146)
College and above (Graduate included)	0.921	0.929	0.932
	(0.151)	(0.153)	(0.153)
Area dummies	YES	YES	YES
Constant cut1	0.0791 ***	0.0804 ***	0.0801 ***
	(0.0286)	(0.0291)	(0.0290)
Constant cut2	0.807	0.820	0.818
	(0.248)	(0.252)	(0.252)
Constant cut3	6.178 ***	6.287 ***	6.267 ***
	(1.880)	(1.913)	(1.908)
Constant cut4	79.25 ***	80.87 ***	80.60 ***
	(24.48)	(24.98)	(24.91)
Observations	6267	6267	6267
Pseudo R-squared	0.100	0.101	0.101

Note: (1) The results are sample-weight adjusted. (2) Standard error reported in parentheses. *** p < 0.01, ** p < 0.05, * p < 0.1. 3) Though not reported here, area fixed effects are controlled in each regression.

reports the results of the basic model of ordered logistic regression analysis, testing the association between SRH and religious affiliations. Column (1) reports the results analyzing the basic factors of reporting happiness among the respondents when no religiosity variables are included. It is noticeable that the odds of reporting a higher level of happiness among rural residents were 0.718 of (or 28.2% lower than) the odds of urban residents (p < 0.05), while the odds of a town resident were insignificant. Similar patterns of the odds ratios between urban and rural residents exist in Column (2) and (3) when religiosity variables are included. Column (2) indicates that the odds of reporting a higher level of happiness among people with a religious affiliation were 1.227 of (or 22.7% higher than) the odds of those who reported to have no religion (p < 0.01). Column (3) reports the results when the specific religious affiliations, Buddhism and Protestantism, were disentangled. The odds ratios of Buddhism and Protestantism were 1.175 (p < 0.05) and 1.399 (p < 0.05) respectively.

Table 3. The association between religiosity/religious practices and happiness (ordered logistic model, odds ratios reported).

	(1)	(2)	(3)
Variables	Happiness	Happiness	Happiness
Religion being very important (Yes)	1.630 ***
	(0.254)
Regular religious attendance (Yes)		1.292 *
		(0.197)
Ever practice praying (Yes)			1.192 **
			(0.0969)
Town resident	0.994	0.999	1.004
	(0.163)	(0.164)	(0.165)
Rural resident	0.724 **	0.725 **	0.721 **
	(0.0990)	(0.0990)	(0.0985)
Observations	6267	6267	6267
Pseudo R-squared	0.101	0.100	0.100

Note: (1) The results are sample-weight adjusted. (2) Standard error reported in parentheses. *** p < 0.01, ** p < 0.05, * p < 0.1. (3) Not reported; however, all regressions have controlled demographic characteristics, socioeconomic feature variables, and area fixed effects that are the same as shown in Table 2. (4) Four constant cuts were included in the ordinary logistic regressions, though not reported in the table above.

reports the results examining how religiosity and religious practices are significantly associated with a higher level of happiness. “Religion being very important” has a highly significant association with a higher level of happiness (odds ratio = 1.630, p < 0.01). Other control variables are the same as in the basic model reported in Table 2.

Table 4. The association between happiness and religious affiliations among urban, town, or rural residents (ordered logistic model, odds ratios reported).

	(1)		(2)
Variables	Happiness		Happiness
Urban without religion	Ref.	Urban without religion + other affiliations	Ref.
Religious	0.983	Buddhist	1.015
urban	(0.109)	urban	(0.111)
Religious	1.377 **	Buddhist	1.390 **
town	(0.186)	town	(0.188)
Religious	1.646 ***	Buddhist	1.324 *
rural	(0.243)	rural	(0.191)
		Protestant	1.004
		urban	(0.276)
		Protestant	2.256 **
		town	(0.815)
		Protestant	1.709 *
		rural	(0.525)
Town	0.956	Town	0.950
	(0.159)		(0.158)
Rural	0.663 ***	Rural	0.686 ***
	(0.0928)		(0.0964)
Observations	6267		6267
Pseudo R-squared	0.101		0.101

Note: (1) The results are sample-weight adjusted. (2) Standard error reported in parentheses. *** p < 0.01, ** p < 0.05, * p < 0.1. (3) Not reported; however, all regressions have controlled demographic characteristics, socioeconomic feature variables, and area fixed effects that are the same as shown in Table 2. (4) Four constant cuts were included in the ordinary logistic regressions, though not reported in the table above.

reports how the association of religiosity and happiness varied among respondent groups in urban, town, and rural residents. The base group includes the urban residents without religion. Despite affiliations with Buddhism and Protestantism, no significantly higher level of self-reported happiness was identified among urban respondents with religious beliefs. However, rural respondents with religious beliefs had a significantly much higher likelihood of reporting a higher level of happiness (odds ratio = 1.646, p < 0.01). The effects were significant among both rural Buddhists (odds ratio = 1.324, p < 0.1) and rural Protestants (odds ratio = 1.709, p < 0.1).

We further did a robustness check by including interaction terms of residential statuses and religious practice indicators such as “religion very important”, “regular religious attendance”, and “ever praying”. The results are reported in

Table 5. The association between religiosity/religious practices and happiness among various residential statuses (ordered logistic model, odds ratios reported).

	(1)		(2)		(3)
Variables	Happiness		Happiness		Happiness
Religion not important (Ref.)
Urban residency X Religion very important	0.855	Urban residency X Regular relig. attend	0.918	Urban residency X ever praying	1.076
	(0.229)		(0.228)		(0.132)
Town residency X Religion very important	2.336 ***	Town residency X Regular relig. attend	1.418	Town residency X ever praying	1.244
	(0.624)		(0.353)		(0.198)
Rural residency X Religion very important	1.916 **	Rural residency X Regular relig. attend	2.312 **	Rural residency X ever praying	1.517 **
	(0.613)		(0.858)		(0.297)
Observations	6267	Observations	6267	Observations	6267
Pseudo R-squared	0.101	Pseudo R-squared	0.101	Pseudo R-squared	0.101

Note: (1) The results are sample-weight adjusted. (2) Standard error reported in parentheses. *** p < 0.01, ** p < 0.05. (3) Not reported; however, all regressions have controlled demographic characteristics, socioeconomic feature variables and area fixed effects that are same as shown in Table 2. (4) Four constant cuts were included in the ordinary logistic regressions, though not reported in the table above.

and the base group refers to those considering religion as not being important. Again, our results indicate that among the respondents with urban residency, none of the religiosity or religious practice activities were associated with a higher level of happiness. Among the town respondents, only the odds ratio of “religion very important” was as large as 2.336 (p < 0.01). However, highly significant and strong associations were again identified among rural religious respondents.

6. Discussion and Conclusions

General association. The results of this study suggest that religious activities (measured by self-identified religious affiliation, religious attendance, or practice) are significantly associated with a higher level of happiness among respondents in China, despite an atheistic social environment and regulations. These findings are generally consistent with the findings in Western societies where religion is a social norm, as well as in peer studies using Chinese national representative samples (Chen and Williams 2016; He et al. 2016; Lu and Zhang 2016; Zhang et al. 2016; Lu and Gao 2017).

Urban versus rural. We found no significant association between religious activities and happiness among urban respondents, but found a strong positive association among rural respondents generally. This finding is consistent with our prediction based on the secularization hypothesis of religion (Norris and Inglehart 2011). Several specific working channels of religion may be considered under the context of China:

(1)

The social network effect of religion could be applied (Lim and Putnam 2010; Krause 2003). An individual’s social network can provide social support, social identification, and social capital to him or her, hence helping to maintain psychological well-being (Halbesleben 2006; Hobfoll et al. 1990). Urban residents in China, especially those in central cities, enjoy a much more advanced social network and opportunities than rural residents (Meng and Chen 2014; Guang et al. 2010). Therefore, the marginal benefits of a social network brought by religious affiliation could be insignificant among urban residents. Meanwhile, rural residents develop their own social relationships based on a kinship network (Steinhardt 2012; Xu et al. 2010) within a narrow range and with limited resources (Meng and Chen 2014; Kipnis 1997). Therefore, religious involvement may help to break the limitation of a kinship network, providing new personal networks and social ties among rural residents (Lim and Putnam 2010).

(2)

Stress coping skills could be obtained. Rural residents in China have a low level of education and few continuous education opportunities, but through religious participation and social activities they might learn better coping skills and psychological supports (Kesebir and Diener 2008). A religious rural resident might be more likely to be satisfied with a materialistically simple life advocated by religious belief, even though the material resources are not comparable to those of an urban resident. Coping skills may be especially helpful among rural residents with chronic diseases, alleviating the psychological burdens of disease.

(3)

Promoting believers’ self-worth and identification (Haslam et al. 2009; Krause and Wulff 2005). Since the social statuses of rural residents in China re generally considered to be at a lower level in the social hierarchy of China, and individuals have long suffered social discrimination (Davey et al. 2009), religion—either Buddhism or Christianity—could help to promote self-worth and self-esteem among rural followers by advocating social equality, inner peace, and forgiveness.

Comparison with the findings of existing studies. Utilizing the same SLSC dataset and logistic models, Lu and Gao (2017) provided the closest existing study to this one. In Lu and Gao (2017), religious identity is measured by whether respondents self-identify as believers of any of the five officially recognized religions, and the authors find that having a religious identity in general is not significantly associated with happiness. This result contradicts our findings, that people identifying as Protestants or Buddhists are more likely to report a higher level of happiness. Despite using the same dataset, the data management in our study is different from Lu and Gao (2017) in several ways. First, we dropped the observations if a respondent was marked as “Had difficulty understanding the survey questions”. Second, we included area fixed effects in all regressions. It is well known that there are very large differences in economic development levels and diversities in the local social cultures among nationwide areas in China.

7. Implications

In each modern society there are rising demands for better health. In China in particular, there is a trend of residents pursuing higher levels of happiness or peak performance of subjective well-being after their basic living needs have been met. In a rapidly ageing society like China, due to the challenges from chronic disease burdens, psychological and/or spiritual needs or coping skills are also in high demanded. Religion can transcend the cognition of patients through the provision of hope, meaning, and life purpose, thus mitigating the difficulty and psychic discomfort of uncertainty regarding the outcomes of certain illnesses (Mueller et al. 2001).

It is expected that the socioeconomic inequality between the urban and rural areas in China will continue to exist. Since rural residents in China can obtain a significantly larger positive effect of happiness enhancement through religious affiliation, it will be necessary for Chinese society to allow free spaces to engage in religious activities, especially for the rural residents, before the social welfare in rural areas can be improved sufficiently.

8. Limitations and Future Study

First, there was a small sample issue among the religious respondents. Despite that the final sample containing 6267 observations, only about 20% of the respondents reported “religious affiliation” and only about 2.41% reported as Christian. The observation numbers of the rural religious believers were also relatively small.

Second, the SLSC dataset contained no information about respondents’ personality characteristics or about the inclinations or motivations of a respondent to practice religion. It provided no information about lifestyle, such as eating habits, risky health behaviors (e.g., drinking or smoking), or exercise. Therefore, the results estimated may be subject to potential unobservable omitted variable biases.

Third, like many peer studies using observatory data to analyze religion and happiness, our analysis also was limited to an associative relationship, rather than being determined by causal effects. There are concerns about the endogeneity or reverse causal effect between religiosity and self-reported happiness (Chiswick and Mirtcheva 2013).

Fourth, there may have be measurement errors and nonrandom missing values in the dataset. Some respondents may have tended not to report their religious beliefs under the general atheistic social environment in China where religion is not a social norm (Lu and Zhang 2016; Wenzel-Teuber 2017), or people may not have had a correct understanding about the terminology “religion” (Wei and Liu 2013). Additionally, some scholars have argued that the varied conclusions of the relationship between religion and happiness depend on the measurement of happiness. Therefore, future studies should concern social background and develop a more highly valid and reliable measurement of religion and happiness for people in a Chinese social context.