Abstract
Prior national studies on religiosity and self-rated health in India have focused solely on older adults using a single data source. Additionally, religion is multidimensional; yet prior research has used limited measures of religiosity. There is a need to expand these measures and examine the general adult population using nationally representative data. The authors investigate the association between five dimensions of religiosity—religious attendance, prayer, scripture reading, religious media consumption, and religious salience—and self-rated health among adults in India. Data were drawn from the Pew Research Center’s 2019–2020 India Survey (n = 27,574), a nationally representative sample of adults 18 years and older. Logistic regression was conducted to assess independent associations between religiosity measures and poor self-rated health, adjusting for covariates. In adjusted models, higher levels of religious attendance and religious media consumption were associated with lower odds of poor self-rated health. Prayer and religious salience were associated with higher odds of poor self-rated health. Scripture reading had a null association. Religious attendance and religious media consumption may be protective social determinants of health for adults in India. Religiosity may be highly valued for those coping with poor health, especially in the forms of religious salience and prayer.
Religion bears on numerous aspects of life, from politics to demographic behavior. Growing evidence also suggests that religion can be a social determinant of health, with various aspects of religiosity driving physical and mental well-being. Research has documented positive associations between religious involvement and various health outcomes, including greater overall physical well-being, lower disease burden, less chronic pain, lower levels of depressive symptoms, less smoking and alcohol use, and longer life expectancy (Hill et al. 2017; Idler 2014; Kent, Bradshaw, and Uecker 2018; Koenig, Peteet, and VanderWeele 2024; Upenieks and Schafer 2020). Dating back to Durkheim (2005 [1897]) and Simmel (1997), one explanation is that religion is influential through groups and social networks. Religious involvement is theorized to affect health because of social ties and interaction between like-minded people. A complementary explanation is that religious practices, especially private practices, can foster meaning and existential certainty in an unpredictable world. Drawing on Mead (1934), Pollner (1989) theorizes that individuals may use such practices (e.g., prayer) in constructing divine relations that provide comfort and direction. Similarly, these practices bolster plausibility structures (Berger 1967) that can offer anomie-reducing order and meaning to life, thus protecting well-being and health. Yet, religiosity is more than group versus private practices. A long line of theory in the scientific study of religion recognizes that religion is a multidimensional and complex phenomenon (Idler et al. 2003; Stark and Glock 1968), and researchers remain without a consensus on which dimensions of religiosity are most important for health.
Within the study of religion and health, another unresolved issue has been the provincial nature of the field, with a growing number of scholars voicing concern in recent years (Zimmer et al. 2016). The vast majority of research has studied populations in the Global North. Thus, there is less certainty around whether religion-health findings hold in Global South societies. More study is needed, in part, because such societies constitute most of the world’s population and tend to be culturally non-Western, economically disadvantaged, and are home to most of the world’s non-Christians. An emerging body of population-based research has begun to address these issues by examining links between religiosity and health cross-nationally, including in an increasing number of societies in the Global South (Bradshaw et al. 2025; Zimmer et al. 2019). An important addition to this emerging Global South research focuses on India (e.g., Roy et al. 2024). India is a lower middle income emerging economy, non-Western, predominantly non-Chrisitan 1 (and predominantly non-Abrahamic), the world’s largest democracy, birthplace and home to multiple large religions, and the world’s most populous country. India is a crucial case.
The present study builds upon this line of research on religiosity and health in India, focusing on self-rated health, and advances the extant literature in multiple ways. First, prior national studies have solely studied older adults using a single data source (Barman et al. 2023; Roy et al. 2024). The present study contributes by using different data that are more recent and comprise a large, nationally representative sample of the general adult population 18 years and older. Second, prior national studies have also been limited in religiosity measurement. The present study uses two measures similar to those used in prior work (religious attendance and prayer frequency) but then extends prior work by incorporating three additional religiosity measures (scripture reading, religious media consumption, and religious salience), thereby providing a more complete picture of religiosity’s links to self-rated health. Additionally, by expanding the multidimensionality of measurement, the present study enhances the ability to isolate the independent association of each form of religiosity with health. Several religiosity measures used are also tailored to respondents’ religious affiliation, thus enhancing contextual relevance.
Background
Recent decades have seen growing attention to the relationship between religion and health. Although a minority of studies has shown negative effects for some aspects of religion and some health outcomes (Hill and Cobb 2011), religiosity has generally been associated with good self-rated health, lower morbidity across a variety of outcomes, and longer life in this body of research (Koenig et al. 2024). For example, prior research has repeatedly shown that religious participation is connected to a variety of mental health outcomes including lower levels of depression, anxiety, and psychological distress symptoms (Baker, Stroope, and Walker 2018; Hill, Burdette, and Idler 2011; Koenig et al. 2024; Stroope et al. 2022). Religious participation has also been linked to various physical health outcomes such as lower odds of hypertension, less chronic pain, and favorable health behavior outcomes (Chen et al. 2020; Gillum and Ingram 2006; Stroope et al. 2025; Rippentrop 2005; Upenieks and Schafer 2020).
Religious involvement is theorized to shape health among adherents in several ways, including through social integration, psychological resources, religious teachings, regulation of behavior, tangible assistance, and by shaping beliefs and norms (Furseth 2017; Green and Elliott 2010; Koenig et al. 2024; Kurtz 2016). As part of social integration, the role of social support is useful in understanding the mechanisms underlying the link between religiosity and health. Religious groups can provide important socioemotional support for their members and can offer companionship, comfort, and sense of belonging (Ellison and Levin 1998; Stroope 2011). These groups can bolster coping mechanisms and tangible assistance in times of need aligning with theological imperatives of compassion and kindness emphasized in many religious traditions (Ellison and Levin 1998; Putnam, Campbell, and Garrett 2012; Stroope, Rackin, and Froese 2021). Religious social integration may also operate through the subjective perceptions of support, which encompasses supportive relationships, acceptance, and anticipated support (Green and Elliott 2010). Religious beliefs, teachings, and accompanying social regulation may shape an individual’s attitudes toward health and health-related behaviors (e.g., alcohol use) (Basu and Dutta 2007; Dutta-Bergman 2004; Ellison and Levin 1998). It is also important to note, however, that the vast majority of this research has been conducted in Global North societies. There is a need for more studies in Global South contexts.
The Indian Context
India is an important case. It is an emerging economy, the world’s most populous country, predominantly non-Christian (and non-Abrahamic) in religious culture, and is the birthplace of multiple popular religions, including Hinduism (79.8 percent of India’s population), Sikhism (1.7 percent), Buddhism (0.7 percent), Sarnaism (0.4 percent), and Jainism (0.4 percent) among others (Pew Research Center 2021). It is also home to large populations of Muslims (14.2 percent of India’s population) and Christians (2.3 percent) (Pew Research Center 2021). Despite rapid economic change, India’s population shows few signs of losing its religion. For instance, both the Indian census and recent survey data find virtually no growth in the small share of people who claim no religious identity (Pew Research Center 2021). By several standard measures, Indians are highly religious, with a majority praying daily (60 percent), more than two thirds visiting a house of worship at least monthly (71 percent), at least three quarters reporting knowing a great deal about their religion, and an overwhelming share saying religion is very important in their lives (84 percent) (Pew Research Center 2021). Religion has been found to have considerable importance for social boundaries and individual behaviors in India. It plays a crucial role in determining social interactions across caste lines, marriage patterns, dietary practices, and political affiliations, among other domains (Evans and Sahgal 2021; Mickevičienė 2003; Robinson 2004).
Health and Religion in India
Health is multidimensional, widely considered to encompass physical, mental and social functioning aspects of health, all of which have long been measured using subjective self-report. Physical and social functioning dimensions of health have also long been measured using objective measures (e.g., blood pressure and observed activities of daily living). Research has established that subjective and objective indicators of health are distinct yet frequently closely related (Barr 2019). For example, studies have found that overall self-rated health measured as a single item is a reliable and valid indicator of health status in the population and consistently predicts later mortality and objective measures of illness (Ferraro and Farmer 1999; Frankenberg and Jones 2004; Idler and Benyamini 1997; Jylhä 2009). The present study focuses on review of self-rated health research, but also notes other health outcomes where relevant. Previous national research conducted on older adults in India has found that prayer frequency is linked to lower levels of depressive symptoms, lower odds of major depressive disorder, and lower levels of cognitive impairment (Barman et al. 2023; Muhammad 2022; Pengpid and Peltzer 2023). Religious attendance is associated with lower odds of cognitive impairment and lower odds of being depressed (Barman et al. 2023; Muhammad 2022). Some national research has investigated the association between religiosity and overall self-rated health in India, but this prior research has been focused on older adults from a single data source. Specifically, a national study in older adults in India found that the frequency of attending religious services and greater spiritual well-being were associated with better self-rated health (Barman et al. 2023). Similarly, a second study using the same national older adult data found that higher frequency of attending religious services and higher frequency of prayer were associated with lower odds of poor self-rated health (Roy et al. 2024).
Beyond Existing Religion-Health Studies in India
The few national studies on religion and health in India, although pioneering, have also been limited in demographic coverage and religiosity measurement. The present study builds upon these studies and advances the literature on religion and health in India in several ways. First, the two prior national studies that examined links between religiosity and self-rated health in India both relied on a single data source, wave 1 of the Longitudinal Ageing Study in India (2017–2018) (Barman et al. 2023; Roy et al. 2024). Although India is aging at a substantially slower rate than other major Asian societies (e.g., China, South Korea), it is certainly important to understand social determinants of health in Indian older adults. Nevertheless, India is one of the world’s youngest major economies with a median age of 28.8 years in 2025 (Statista 2023). Indeed, a significant share of India’s population is of working age; study of the general adult population, including young adults and those in early midlife, is needed. Furthermore, there are empirical and theoretical reasons to expand beyond studying older adults. Research from other societies suggests that the positive effects of religiosity on self-rated health decrease with age (Zhang 2017). Additionally, rooted in the insights of life-course theory (Elder, Johnson, and Crosnoe 2003), the mechanisms underlying religiosity’s protective health effects are thought to differ in younger adulthood (Chen and VanderWeele 2018; Freeman 2021; Guo and Metcalfe 2019). Additionally, the nature of evidence is such that it accumulates over multiple studies and diverse samples; therefore, using different data sources, especially those with different characteristics, is scientifically critical (Freese and Peterson 2017). Taken together, India’s young demographic profile combined with insights on life course mechanisms suggest that we need to understand links between religion and health, not solely in older adults in a single dataset, but also in other data representative of the general adult population. The present study addresses this lacuna by examining the links between religion and health in a large, nationally representative sample of the general adult population in India.
Second, prior national religion-health studies of older adults in India have differed in and been limited in measurement of religiosity. Barman et al. (2023) used a measure of religious function participation and a general-use measure of spiritual well-being not tailored to the Indian context. Roy et al. (2024) used a measure of prayer frequency, religious service attendance, and meditation frequency, all of which included some tailoring to the overall Indian context, particularly pertaining to Hinduism. These approaches are important because religion is long known to be multidimensional and an important unresolved issue in the field is gaining a fuller understanding of how different dimensions of religion may have distinct effects on health (Idler et al. 2003). Including measures for multiple dimensions is also important for isolating their independent effects. Relying on only one or two measures of religiosity may overlook other aspects of religiosity and thus overstate associations with health compared to these other aspects of religiosity (Idler et al. 2009; Levin 2012). The present study extends the multidimensionality of this prior work by adding three additional measures of religiosity, thereby providing a more complete and encompassing analysis. Furthermore, the present study builds on prior work by using contextually specific measurement. Specifically, the present study is the first to assess religion-health links using items that were not only tailored to the general Indian context but were tailored on the basis of each respondent’s reported religious affiliation (e.g., Muslims were asked about going to the mosque for namaz, not going to a temple or attendance at “religious services”). This approach was applied to all relevant religiosity measures in the study, affording contextual specificity for diverse religious traditions and allowing greater relevance to the local setting.
Facets of Religion and Health
Involvement in religious communal life is an important aspect of being religious that can generate a sense of belonging, shared values, and supportive networks among individuals (Ellison and Levin 1998; Stroope 2011). It can promote positive health behavior and deter unhealthy behavior through normative influence, social sanctions and fostering beneficial health beliefs (Ellison and Levin 1998; Hall, Meador, and Koenig 2008; Hill and Cobb 2011; Krause 2008). The frequency of religious attendance is a key indicator of communal religious involvement and is one of the most frequently measured aspects of religiosity in studies of religion and health. A substantial body of research has found that regular religious attendance is associated with better health and a longer lifespan (Chen et al. 2020; Koenig et al. 2024). Although it is not entirely understood why this relationship exists, some research suggests that social support may be a key factor (Idler 2014). For example, involvement in religious communal settings appears to foster social connections and a sense of belonging, which can lead to better health (Stroope and Baker 2014). On the basis of this theoretical background, we hypothesize that a higher frequency of religious attendance is inversely associated with poor self-rated health (hypothesis 1).
Aside from communal forms of religiosity, a variety of other religious practices can also be consequential for health. These sometimes fall under the heading of “nonorganizational” or “private” religious practices; however, depending on study measurement and cultural context, these practices may or may not always be strictly nonorganizational or private in nature. An example is the frequency of prayer. Prayer has been a perennial form of religiosity practiced across a wide range of religions. It has also been used as a health intervention (Edman and Koon 2000; Masters and Spielmans 2007). When studying the impact of religion on health, studies from a variety of settings have tended to use prayer as an indicator of private religiousness (Ellison and Levin 1998; Doane and Elliot 2016). Although prayer is more widely considered to be oriented toward the private and psychological, its significance as a sociological phenomenon is also apparent. In some cultural contexts (e.g., India), prayer is often practiced socially or publicly (Dutta-Bergman 2004). As such, prayer may be socially integrative and health-promoting (Bradshaw and Kent 2018). However, it is important to note that literature on prayer and health outcomes has not always shown favorable associations. Some studies have found null associations between prayer and various health outcomes, whereas others have reported that more frequent prayer is associated with poorer health (Koenig et al. 2024). The latter finding may be because individuals turn to prayer for comfort and resilience during ill health. However, given the balance of associations in the literature, we hypothesize that higher frequency of prayer is inversely associated with poor self-rated health (hypothesis 2).
Similarly, recent research has also shown that a variety of other practices are associated with health outcomes. For example, reading religious scriptures is related to psychological resilience and better mental health. One plausible mechanism for this relationship is that scripture reading may promote coping strategies and constructive reinterpretations of stressful events. For instance, studies have found that engaging with religious texts mitigates the detrimental impacts of stressors on mental health by encouraging positive religious reappraisals (DeAngelis et al. 2021; Krause, Pargament, and Ironson 2018). Drawing from these insights, we posit that higher frequency of reading religious scriptures is inversely associated with poor self-rated health (hypothesis 3). Consuming religious media, such as watching religious videos or listening to religious podcasts, is another facet of religious involvement that can foster a sense of connection with coreligionists and provide positive emotional and psychological reinforcement associated with well-being (Faverio et al. 2023; Iyer, Larcom, and She 2024). Through popular devotional programs, religious media consumption in India may provide easily accessible spiritual resources and health information (Kumar 2006). Thus, we hypothesize that more consumption of religious media is inversely associated with poor self-rated health (hypothesis 4).
In addition to religious practices, various subjective aspects of religiosity such as an individual’s sense of religious devotion and importance of religion in daily life may also influence health (Benjamins 2004). These aspects of religiosity, which represent an individual’s overall assessment of their religiousness, may affect well-being by providing an enhanced sense of interconnectedness, purpose, and meaning in life (Hill et al. 2000; Park, Edmondson, and Hale-Smith 2013). Although less researched than some other religiosity measures, religious salience, a measure of how significant religion is in an individual’s life, is found to have a favorable associations with mental health, physical health, health behaviors, and overall self-rated health (Al-Kandari 2011; Kobayashi et al. 2015; Montague et al. 2003; Nicholson, Rose, and Bobak 2010). And given the frequent interaction between religion and facets of daily life and social identity in India, religious salience may be particularly germane to the religion-health link in India (Muhammad 2022; Sen, Wagner, and Howarth 2014). On the basis of this background, we hypothesize that greater salience of religion is inversely associated with poor self-rated health (hypothesis 5).
A minority of studies have also found that religiosity is not always associated with better health (Hill and Cobb 2011; Pruyser 1978; Rippentrop 2005; Schafer 1997). Individuals may resort to religion for comfort because of ongoing health problems, which can result in increased participation in practices such as prayer (Doane and Elliot 2016; Idler 1995; Levin 1989). Such religious coping following health problems would be consistent with the negative association between religiosity and health seen in some studies. Another possibility is that frequent exposure to collective religious settings is not always conducive to positive feelings and supportive relationships. For example, a substantial minority of research has shown that frequent religious attendance is related to higher anxiety levels (Koenig, Al-Zaben, and VanderWeele 2020; Schieman, Bierman, and Ellison 2013; Stroope et al. 2022), and congregational neglect has been found to help explain this connection (Stroope et al. 2024). Additionally, experiencing criticism from members of religious congregations has been connected to increased levels of anger and anxiety (Kent et al. 2020). Despite these potential negative effects, prior studies most commonly indicate positive relationships between religion and health outcomes. This modal pattern in the literature guides hypothesized associations in the present study. In the analysis that follows, we examine the multidimensionality of religion as it relates to health in the general adult population in India by measuring five religiosity measures, including several tailored to the Indian context, assessing their independent relations to self-rated health.
Data and Methods
The data for this study came from the Pew Research Center’s 2019–2020 India Survey (Pew Research Center 2021), conducted between November 2019 and March 2020. A total of 29,999 Indian adults were surveyed using face-to-face surveys, with 22,975 identifying as Hindu, 3,336 identifying as Muslim, 1,782 identifying as Sikh, 1,011 identifying as Christian, 719 identifying as Buddhist, 109 identifying as Jain, and 67 identifying as belonging to another religion or as religiously unaffiliated. The survey was conducted using a multistage, stratified cluster sampling method and was designed to be nationally representative. Participants were identified using a probability-based sample design to ensure that the most widely known religious groups in India—Hindus, Muslims, Christians, Sikhs, Buddhists, and Jains—as well as all major regional zones were adequately represented. The survey reached 98 percent of Indians 18 years and older, with an 86 percent response rate. The questionnaire to conduct this study was developed in English and then translated into 16 regional languages and verified by qualified linguists proficient in regional dialects. Pew’s India Survey is the most comprehensive survey of religion in India to date. It is unique in its scale, scope, oversampling of minority groups, detailed data collection, culturally appropriate questionnaire development, and face-to-face interviews in multiple languages. These features help overcome the limitations of previous cross-sectional datasets, as they provide more comprehensive and relevant data on the role of religion in the lives of Indian adults, thereby helping address limitations in previous studies of religion and health in India.
Dependent Variable
The dependent variable in this study is a binary measure of self-rated health. Participants were asked, “Nowadays, would you say that your health is . . .?” (response options were “excellent,” “good,” “fair,” and “poor”). We binarized self-rated health by coding those reporting good or excellent health as 0 and those reporting fair or poor health as 1. For simplicity, we label this variable poor self-rated health. We elected to binarize self-rated health because of failure to pass the Brant test (both omnibus and for key predictors) and after probing multinomial logit model results as described under “Analytic Method.” Additionally, dichotomizing as fair or poor health is in keeping with much prior research using self-rated health in India and other contexts, thus making our results more directly comparable (e.g., Akhtar, Saikia, and Muhammad 2023; Davis et al. 2017; Do, Frank, and Iceland 2017; Saha et al. 2022; Roy et al. 2024; Stroope, Kroeger, and Fan 2021). Self-rated health is frequently used in demographic and population health surveys to collect information on respondents’ overall health. Prior studies have supported its validity as a reliable measure of assessing health status, including in India, and as a consistent predictor of current illness and future mortality and morbidity (Cullati et al. 2018; Ferraro and Farmer 1999; Frankenberg and Jones 2004; Idler and Benyamini 1997; Jylhä 2009; Subramanian et al. 2009).
Independent Variables
The key independent variables used to measure the religiosity of respondents in this study include frequency of religious attendance, salience of religion, frequency of prayer, reading religious texts, and watching religious programs. The respondents of the survey were asked about their religious attendance habits. Most survey questions were tailored to the respondent’s religious affiliation as follows: for Hindus, Buddhists, and Jains, “How often do you go to the temple?” for Muslims, “On average, how often do you attend the mosque for namaz?” for Sikhs, “How often do you go to the gurdwara?” for Christians, “Aside from weddings and funerals, how often do you attend religious services?” and for all other religious affiliations, “How often do you attend religious services?” Responses ranged from “never” (1) to “every day” (7). Religious salience measured how personally significant and important individuals perceive their own religious beliefs and practices to be. The question asked was “How important is religion in your life?” Response options were “not at all important” (1), “not too important” (2), “don’t know” (3), “somewhat important” (4), and “very important” (5). The survey question for measuring frequency of prayer was also tailored to religious affiliation. Non-Muslim respondents were asked, “Do you pray?” and Muslim respondents were asked, “Do you offer namaz?” Responses ranged from “never” (1) to “several times a day” (7). Additionally, frequency of watching religious programs was asked as “How often do you watch religious programs or serials if at all?” The responses ranged from “never” (1) to “every day” (7). The respondents of the survey were asked about their frequency of reading religious texts, also tailored to religious affiliation. Hindu respondents were asked, “How often do you read or listen to recitations of religious books, for example, the Bhagavad Gita?” Muslims were asked, “How often do you read or listen to recitations of religious books, for example, the Quran?” Christians were asked, “How often do you read or listen to recitations of religious books, for example, the Bible?” Sikhs were asked, “How often do you read or listen to recitations of religious books, for example, the Guru Granth Sahib?” Buddhists were asked, “How often do you read or listen to recitations of religious books, for example, the Sutras?” Other religionists were asked, “How often do you read or listen to the sacred text of your religion?” Responses ranged from “never” (1) to “every day” (7).
Covariates were also included that have been previously identified as potentially confounding and that were available in the Pew India Survey (Barman et al. 2023; Krause 2010): age, gender, region, rural, religious affiliation, caste, education, marital status, household size, joint or nuclear family, and number of children in the household. Age was measured in years as a categorical variable: 18 to 25, 26 to 34, 35 to 44, 45 to 59, and 60 years or older. Historically, India’s caste system has divided people into social groups, and although caste is mainly associated with Hinduism, it is also present in other religions in India (Ali 2002; Blunch and Gupta 2014). In the present study, caste was coded into five categories: forward caste, other general caste, other backward class, scheduled caste, and scheduled tribe. Educational attainment was also a categorical variable with six categories ranging from preprimary to highest tertiary. 2
Marital status of respondents was a categorical variable: married or living with a partner; divorced, separated, or widowed; and unmarried. Region was a categorical variable: Northeast, North, Central, East, West, and South. Additionally, an indicator variable for rural was included (1 = yes, 0 = else). Some household characteristics were as follows: whether the respondent lives in a joint family (1 = yes, 0 = else), household size, and number of children in the household (ranging from no children [1] to three or more children [4]). Religious tradition was a categorical variable: Hindu, Muslim, Christian, Sikh, Buddhist, Jain, and other or no religion.
Analytic Method
After excluding data missing on self-rated health (n = 38 [0.1 percent]), data were missing for some predictor variables (7.9 percent), resulting in an analytic sample of 27,574. Following compilation of summary statistics, the associations between religiosity measures and poor self-rated health were examined using multiple regression. We estimated two regression models. The first model included only the focal religiosity variables. The second model entered sociodemographic covariates. All analyses applied the recommended sampling weight and were conducted using Stata 15. An ordered logit model was not used, because a Brant test showed that an ordered logit model violated the proportional odds assumption (p < .001, omnibus Brant test). Specifically, religious attendance, scripture reading, and religious media consumption were all related to self-rated health in a nonordered fashion (p < .01, variable-specific Brant tests). Ancillary analyses treating self-rated health as a four-category nominal outcome supported dichotomizing self-rated health as fair or poor health = 1 and good or excellent health = 0. Specifically, in a multinomial logit model, the pattern of results indicated that predictor variable effects were not significant in distinguishing between poor and fair health, however when poor or fair health were treated as the base category, focal religiosity variables significantly predicted differences between poor or fair health on the one hand with good or excellent health on the other hand. As multinomial logit models entail many contrasts and are complex to present and interpret, we elected to present binary logit results which more parsimoniously capture the overall results.
Results
Summary statistics for study variables are shown in Table 1 and a correlation matrix of poor self-rated health and religiosity variables is presented in the Supplementary Materials (Table S1). Table 2 presents the results of the logistic regression models. The direction of effects and statistical significance of religiosity variables were generally consistent between model 1 (without sociodemographic controls) and model 2 (with sociodemographic controls). However, frequency of reading scripture lost statistical significance in model 2. In model 2, all religiosity variables were significantly associated with self-rated health (except scripture reading). 3 Furthermore, comparison of information criteria between model 2 and a model where a given religiosity variable was omitted indicated very strong evidence (change in Akaike information criterion and change in Bayesian information criterion were both >10) for adding each of the significant religiosity variables to the model. The specific reductions in Akaike information criterion and Bayesian information criterion when adding these variables were as follows: religious attendance, 21.677 and 13.452; prayer, 52.432 and 44.208; religious media consumption, 111.854 and 103.629; and religious salience, 18.185 and 9.961. The direction of the individual associations, however, did not always align with the hypothesized direction. Higher frequency of religious attendance was associated with lower odds of poor self-rated health, supporting hypothesis 1. Contrary to hypothesis 2, higher frequency of prayer was associated with higher odds of poor self-rated health. Hypotheses 3 was ultimately not supported; though the effect of scripture reading was significant in model 1, no statistically significant effect of frequency of reading was observed in model 2. In support of hypothesis 4, higher levels of religious media consumption were associated with lower odds of poor self-rated health. Finally, higher religious salience was linked to higher odds poor self-rated health (hypothesis 5 unsupported).
Summary Statistics.
Note: Data are weighted.
Logistic Regression of Poor Self-Rated Health (n = 27,574).
Note: Data are weighted. b = unstandardized coefficient; β = standardized coefficient; Δ = average marginal effect.
p < .05. **p < .01. ***p < .001.
Comparing standardized coefficients from model 2 is one way to gauge the relative magnitude of the effect of each of the significant religiosity measures. In order of magnitude, these coefficients were as follows: religious media consumption, β = −0.337; prayer, β = 0.242; religious attendance, β = −0.156; and religious salience, β = 0.139. Alternatively, average marginal effects (AMEs) compute, on average, the effect on the outcome of a change in a focal predictor across every observation in the data. We computed AMEs for a 1-SD increase in each predictor. Specifically, +1 SD in religious attendance frequency was associated with a 1.5 percentage point decrease in the predicted probability of reporting poor self-rated health (Δ = –0.015). A +1-SD difference in prayer frequency was linked to a 2.3 percentage point increase in the probability of poor health (Δ = 0.023). A +1-SD difference in religious media consumption was associated with a 3.2 percentage point decrease in the probability of poor health (Δ = –0.032). Last, a +1-SD difference in religious salience was linked to a 1.3 percentage point increase in the probability of poor health (Δ = 0.013).
We also conducted tests for differences between pairs of marginal effects, which assessed whether the difference in the AMEs of two predictors is statistically significantly different from zero. Six pairwise comparisons were conducted to assess whether the absolute values of the AMEs (regardless of direction) differed significantly. We found that none of the comparisons reached conventional levels of statistical significance. For example, the absolute values of the AMEs for religious attendance and religious media do not significantly differ. Nor do those of religious media and prayer. This suggests that, though the direction of effects differ between religious attendance and religious media on the one hand (negative) and prayer and salience on the other hand (positive), the absolute size of the marginal effects for any of these variables do not significantly differ.
Discussion
The aim of the present study was to advance our understanding of the relationship between religiosity and health. Although prior research on religiosity and health has mostly concentrated on Global North contexts, emerging research has contributed analysis from Global South contexts, including a handful of national studies on older adults in India, the world’s most populous nation and a key non-Western, economically emerging, and majority non-Christian case (e.g., Roy et al. 2024). The present study contributed to our understanding of overall self-rated health in this research literature. We went beyond prior work that studied older adults by using a large, nationally representative sample of the general adult population 18 years of age and older. We also extended past work by assessing five measures of religiosity, three novel to national research on religion and self-rated health in India, thus, offering a more encompassing approach that takes the multidimensionality of religion seriously and better isolates independent contributions of different forms of religiosity. Additionally, we advanced prior work by using religiosity measures tailored to respondents’ specific religious traditions, allowing greater relevance to the diverse expressions of religiosity in the Indian context.
We hypothesized that all five religiosity measures would be associated with lower odds of poor health and showed that four religiosity measures significantly associated with health in adjusted models. We found that religious attendance and religious media consumption were associated with lower odds of poor health, whereas frequency of prayer and religious salience were associated with higher odds of poor self-rated health. Reading scriptures did not have a statistically significant effect on health in adjusted models. We contributed to extant work by assessing religious attendance and prayer in the general adult population and found that results for prayer differed in direction from past work. Our results contributed novel evidence for the association between religious media consumption and the lower odds of poor health, religious salience and higher odds of poor health, and the null effects of scripture reading.
In keeping with hypothesized expectations, our results showed that a higher frequency of religious attendance was associated with a lower likelihood of poor self-rated health. Participating in religious community, or the frequency of attending religious services, is the religiosity indicator most consistently associated with better health and longevity in prior research (Chen et al. 2020; Hummer et al. 1999; VanderWeele, Jackson, and Li 2016). In previous research on older adults in India, religious attendance was found to be associated with better self-rated health (or lower odds of poor self-rated health), lower levels of cognitive impairment, and lower levels of depressive symptoms and diagnosis (Barman et al. 2023; Muhammad 2022; Pengpid and Peltzer 2023; Roy et al. 2024). The present study’s finding is in keeping with these previous gerontological findings and extends evidence of the association between religious community and overall self-rated health in India to the general adult population.
Our findings showed a relationship between prayer frequency and self-rated health contrary to that hypothesized. Individuals who prayed more frequently were more likely to report poor self-rated health. This finding differed from the previous research on older adults in India which found prayer to be associated with better self-rated health, lower levels of depressive symptoms, and lower levels of cognitive impairment (Chokkanathan 2013; Muhammad 2022; Pengpid and Peltzer 2023; Roy et al. 2024). In India, prayer is commonly used to lessen or eliminate unfavorable circumstances (e.g., poor health) (Andrade and Radhakrishnan 2009; Babula 2023). As such, it may be that unfavorable health conditions lead to a rise in prayer frequency as a coping behavior. The complex relationship could be influenced by the timing and circumstances of prayer, such as intensified prayer during illness. The divergence of our findings from prior work may also be influenced by methodological differences across studies, such as sample demographics and the measurement of prayer. Indeed, despite research from a variety of settings suggesting a positive association between prayer frequency and health, ours is not the first study to find that prayer frequency is associated with poor self-rated health. Studies from other settings have also found patterns consistent with prayer as coping response, an inverse relationship between prayer and health (Cowden et al. 2022; Nicholson, Rose, and Bobak 2009). This suggests that higher prayer frequency may often be an act of seeking solace following health problems. It is possible that the frequency of prayer could be both a potential contributor to health and a reaction to poor health. This bidirectional relationship highlights the need for caution when interpreting the role of prayer in health, particularly in cross-sectional studies where temporal order cannot be definitively established. More study is needed to understand these mixed findings regarding prayer.
Contrary to expectations, we found that higher frequency of scripture reading was unrelated to self-rated health after adjusting for sociodemographics. In past research from the United States, scripture reading has been shown to have varying associations with mental health (DeAngelis et al. 2021; Krause and Pargament 2018). This research theorizes that scripture reading provides readers with coping resources in the form of narratives that encourage benevolent reappraisals of challenging life circumstances, thus bolstering mental health and other potential health outcomes. Until now, no national research has examined the health effects of scripture reading in India. Although some studies from other contexts have shown that scripture reading is adversely related to health (Krause and Pargament 2018), more research is required across a variety of cultural contexts. The present study is the first known analysis to assess this issue in a national sample of India and concluded that scripture reading is not robustly associated with health.
Consistent with expectations, we found that higher frequency of religious media consumption was related to lower odds of poor self-rated health. Studies of religious media consumption have focused on the potential effects of religious media at the intersection of religious involvement and mental health (Stroope, Walker, and Franzen 2017). It is well established that specific media consumption may be driven by psychological gratification and identity formation (Couldry, Livingstone, and Markham 2007; Hjarvard 2011). Religious media consumption such as watching religious television programs, attending online prayer services, or listening to religious podcasts, are relatively modern forms of religious practice. Consuming religious media, though typically thought of as a private practice in the West, may be private or communal in India. The communal nature of religious media was perhaps best observed in the closing of large portions of the economy and simultaneous gatherings around screens as more than 650 million viewers watched the Hindu epic Ramayan from January 1987 to July 1988 in cities and villages across India, the most watched Indian serial of all time (Sagar 2019). Consuming religious media in groups can make it a communally shared and interactive experience, thereby fostering social integration and regulation among the participants. The results from this study advance this relatively underexplored issue and support the hypothesis that consumption of religious media is favorably associated with health in India. Further measurement and study are needed to understand the pathways through which religious media could promote health.
We hypothesized that religious salience would be linked to lower odds of poor self-rated health. We found the opposite association, thus contradicting our hypothesis. Similar to the pattern for prayer, this divergence in findings for salience may be due to people’s heightened importance of religion in the face of suffering. When health problems emerge, people may turn to religion as an important way to cope and find comfort and meaning amid pain. This finding may also indicate a discrepancy between the value individuals place on religion and their actual religious practices. It is important to distinguish between the subjective importance an individual places on religion and an individual’s more objective religious practices (Babula 2023; Benjamins 2006; Idler et al. 2009). In the Indian context, many people value religion; however, valuing religion may not be a reliable indicator of a person’s actual degree of religiosity. This potential disconnect between religious salience and practiced religiosity is an important consideration for understanding the relationship between religiosity and health.
In summary, we found that religious attendance and consumption of religious media were associated with better self-rated health. Prayer and religious salience were associated with worse self-rated health. Scripture reading did not maintain an effect. Religious attendance, typically a public and communal activity, and religious media consumption, which may also involve considerable social interaction, likely contribute to better health through their socially integrative and regulative functions. Conversely, prayer and salience may become more prominent in response to health challenges, suggesting a reverse causal pathway where religiosity serves as a form of coping. Taken together, these findings underscore the importance of distinguishing between different dimensions of religiosity when examining their relationship to health and doing so in a variety of national contexts. Future studies should examine these relationships further to clarify their directionality and underlying mechanisms.
Limitations and Strengths
The reliance on cross-sectional data in this study is one of its limitations, which hinders establishing temporal ordering between religiosity and health. The cross-sectional nature of the study limits the ability to exclude health selection as a contributing factor. This is an important consideration as having poor health may inhibit later religious community participation. For a better understanding of the dynamic interplay between religion and health, future research should use nationally representative longitudinal samples of the general adult population in India. Another limitation of this study is its inability to measure potential mediating mechanisms between religiosity and health, such as social support, social regulation, and sense of meaning, among others. For example, religious attendance can offer opportunities for social interaction, which can foster social ties that lead to social support and a sense of belonging. Similarly, communal prayer settings, scripture reading, and religious media consumption can connect individuals to others and foster shared narratives and meaning-making (Idler 2014; Krause and Pargament 2018; Stroope, Draper, and Whitehead 2013). Future studies should directly assess such mechanisms in the Indian context. Third, we theorized that in the Indian context, prayer may be both private and highly communal, yet the study’s survey question does not distinguish between these. Future research should distinguish between private and communal prayer and assess their independent relations to health. Finally, like much other social science research (e.g., Stroope 2015), the current study relied on a subjective measure of health. Both objective and subjective measures of health are crucial for a complete picture of well-being. Because of the distinctive nature of objective measures such as biomarkers, researchers have increasingly called for the inclusion of biomarkers and other objective measures in studies of the social determinants of health. Nationally representative studies of Indian adults 18 years and older that use both subjective and objective measures are needed. Nevertheless, self-rated health is a reliable and valid indicator of current health status, including in India, and predicts later mortality and morbidity (Ferraro and Farmer 1999; Frankenberg and Jones 2004; Idler and Benyamini 1997; Jylhä 2009; Subramanian et al. 2009).
Conclusion
The aim of this study was to advance understanding of the relationship between religiosity and health in adults in India. It is novel in its breadth of religiosity measurement and unique in assessing religiosity and health in the general adult population in India. The results indicate that religious attendance and religious media consumption warrant consideration as protective social determinants of health. Religion may also be highly valued in coping with flagging health, especially in the form of religious salience and prayer. Considering secularization trends across many parts of the world, India’s enduring religious intensity offers a distinctive and valuable context for examining the relationship between religion and health. Understanding how religiosity shapes health is especially important in a society where religion remains deeply intertwined with everyday life. Importantly, this study joins others in suggesting that religiosity can serve as a social factor germane to health, not just in majority-Christian contexts and societies of the Global North but also in the Global South in a major non-Christian setting such as India, where traditions such as Hinduism predominate. Given India’s status as an emerging economy, religion may play an especially important role in promoting health within a resource-constrained environment, where other resources for health promotion may be limited. Furthermore, with India’s demographic shift toward midlife and older age in the coming decades, understanding social determinants of health, such as religiosity, carries considerable policy relevance. Understanding these influences can help policymakers design culturally sensitive approaches that leverage the strengths of religious communities to improve public health in the world’s most populous nation.
Supplemental Material
sj-docx-1-srd-10.1177_23780231261416210 – Supplemental material for Self-Rated Health among Adults in India: Examining the Role of Religiosity
Supplemental material, sj-docx-1-srd-10.1177_23780231261416210 for Self-Rated Health among Adults in India: Examining the Role of Religiosity by Harleen Sandhu and Samuel Stroope in Socius
Footnotes
Acknowledgements
The authors thank colleagues for their valuable feedback on earlier drafts of this article.
Supplemental Material
Supplemental material for this article is available online.
1
Christians are a small minority population in India and represent 3 percent of this study’s analytic sample. Some regional research has suggested physical health disadvantages for Christians versus other large religious groups in India (e.g., Vijayakumar, Arun, and Kutty 2009).
2
Pew constructed a binary household income variable, but we excluded this variable because of its large number of missing values. Results in a sensitivity analysis including income were substantively unchanged.
3
Variance inflation factors ranged from 1.06 to 1.52, indicating low multicollinearity. Results remained consistent when each religiosity variable was modeled independently, without controlling for other aspects of religiosity.
Author Biographies
References
Supplementary Material
Please find the following supplemental material available below.
For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.
For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.
