Religious Affiliation Is Not Enough: Considering the Religious Practices and Self-Identification of Seniors in Switzerland When Measuring the Links Between Religiosity and Well-Being

Abstract

Previous studies have shown that religion and spirituality can provide resources to cope with aging and impact the well-being of older adults. However, whether religiosity and well-being are linked depends on how they are measured. Moreover, the size of statistically significant effects often remains small or even negligible. In Switzerland, two historical religious communities—Catholic and Protestant—have coexisted for centuries, providing an opportunity for exploring potential differences in religious practices of members of these two communities and their effects on well-being. Two populations were targeted: elderly people engaged in organized volunteering and elderly people receiving home care services. A total of 617 volunteers and 614 home care clients returned complete and valid questionnaires. In addition, semi-structured interviews were conducted with 43 volunteers and 35 home care clients to gain deeper insights into their experiences. Since the data were collected during the COVID-19 pandemic, we also explored how participants experienced the semi-lockdown restrictions. The results show that knowing religious affiliation is not sufficient to predict well-being. The degree of identification with the declared affiliation must also be taken into account. Differences between Protestants and Catholics can be highlighted, especially when considering religious practices in more detail. The impact of the pandemic on well-being is only moderately associated with religiosity.

1. Introduction

As early as 1897, Émile Durkheim highlighted differences between Protestants and Catholics with regard to the preservation of life. In his study on suicide, he compared different European countries. He noted that “in purely Catholic countries, such as Spain, Portugal, Italy, suicide is very little developed, while it is at its maximum in Protestant countries, in Prussia, Saxony, Denmark.” (Durkheim 1897, p. 149). Recognizing that differences between countries could be due to factors other than religious affiliation, he further examined this by comparing different regions within the same country and showed that suicide rates vary proportionally to the rates of Protestants in these regions and inversely to the rates of Catholics. In Switzerland, he also notes that “Catholic cantons account for four to five times fewer suicides than Protestant cantons” (Durkheim 1897, p. 151), and this is true in both French-speaking and German-speaking Switzerland. Durkheim attributes this difference to individualism. Believing that “the only essential difference between Catholicism and Protestantism is that the latter allows for free examination to a much greater extent than the former” (Durkheim 1897, p. 156), he concludes that the Protestant, “more the author of his belief” (Durkheim 1897, p. 157), belongs to “a less strongly integrated Church than the Catholic Church” (Durkheim 1897, p. 159). In other words, it is the degree of social integration that constitutes the determining factor in the suicide rate. Social integration is said to have a protective effect.

A recent Swiss study confirms the observations made by Durkheim. Spoerri et al. (2010) examined data from the 1990 and 2000 national censuses (records of 3,194,911 individuals aged 35 to 94 years) and demonstrated that suicide rates are higher among Protestants: per 100,000 inhabitants, the rates were 19.7 among Catholics, 28.5 among Protestants, and 39.0 among those with no religious affiliation. This suggests that, despite the significant secularization of modern Swiss society and the decreasing doctrinal and social differences between Catholicism and Protestantism, a denomination effect persists, with Catholics exhibiting lower suicide rates than Protestants. In light of these findings, it becomes pertinent to investigate whether such denomination effect also extends to other dimensions closely linked to suicide risk, namely mental health and well-being. This is precisely the focus of the present article: exploring the links between religion, health, and well-being, particularly when comparing Protestants and Catholics. We will begin by providing a brief overview of the influence of religion on health, drawing primarily on recent literature reviews and meta-analyses. We will then examine in more detail the comparison between Protestants and Catholics in the literature before presenting the results of data recently collected in Switzerland, in the canton of Vaud. For interested readers, a book presents the results of various analyses of this data, though it does not include the comparison between Protestants and Catholics (Brandt et al. 2026).

2. Literature Review and Context

The issue addressed in this article is situated in the background of research on the links between religiosity, well-being and health. The literature review highlights how research has shifted from studies investigating links between religiosity, health, and well-being to a greater focus on how these concepts are defined. Indeed, the possibility of asserting the existence of significant links depends largely on how religiosity and well-being are measured. Moreover, the interpretation that can be given to significant links depends on the effect size. This section of the article problematizes these very links before focusing more specifically on studies comparing Protestants and Catholics, and then on the Swiss context.

2.1. Religiosity, Health, and Well-Being

It has become common to claim that religion has a positive impact on well-being (Diener et al. 2011; Geerling and Diener 2020; Hoogeveen et al. 2023; Koenig and Larson 2001; Ugur and Aydın 2023; Yaden et al. 2022). For example, to take only the aging population on which our study focuses, Koenig already listed in 1995 nearly 100 studies on the positive effects of religion/spirituality on health, well-being, and quality of life in older people and terminally ill patients. In the two decades that followed, a large number of studies have confirmed this trend at both the physical and psychological levels. Thus, still with regard to the elderly population, studies have shown many positive effects of religion/spirituality: a decreased risk of stroke (Colantonio et al. 1992), an inverse relationship between religious activity and systolic blood pressure (Koenig et al. 1998); attendance at a religious service at least once a week associated with a lower allostatic load index in older women (Maselko et al. 2007); better morale (less anxiety and a better sense of self-efficacy) (Wassel Zavala et al. 2009); better psychological or psychosocial adjustment to illness, better physical health, reduced interference due to pain, and better quality of life (Davison and Jhangri 2013); improved social function (Wang et al. 2008); a positive influence on self-esteem (Krause 1995; Krause 2003); a higher level of self-efficacy (Levin and Chatters 2008); optimism and life satisfaction (Keyes and Reitzes 2007; Koenig 1995; Krause 2003; Moberg 2005; Turesky and Schultz 2010; Yoon and Lee 2006); an association between high religious faith and spirituality and an increased ability to cope with stress and adversity (Ai et al. 1998; Barusch et al. 1999; Krause and Van Tran 1989); lower levels of anxiety (Cicirelli 2002; Vink et al. 2008); more comfort in times of distress and uncertainty (Ai et al. 1998; Fiske et al. 2009; Fry 2001; Koenig 2006; Levin and Chatters 2008; Powell et al. 2003); and protection against distress at the end of life, particularly by reducing the fear of death (Cicirelli 2002).

Various explanatory mechanisms for the link between spirituality/religiosity and the well-being and health of seniors have been put forward in studies: the promotion of healthy living and prohibitions on the consumption of drugs, alcohol, tobacco, risky sexual behavior, etc. (George et al. 2000); the ability of the religious and spiritual dimension to provide a positive interpretative framework facilitating acceptance (e.g., of illness) and adaptation (Persson et al. 1999); the facilitation of social support (studies show that believers have a larger social network, more social interactions, thus more social support and more satisfaction with this support) (Bartlett et al. 2003); the ability to give meaning (Hill and Pargament 2003); the mediating effect of spirituality against stress and for psychosocial development (Brennan 2002).

This could be supplemented by a review of more recent publications, continuing along the same lines. However, in recent years, meta-analyses, cross-national studies, and a many-analysts approach have highlighted methodological problems regarding the measurement of the links between religion and well-being in general, and not just among older people.

In a cross-national study, Hayward and Elliott (2014) analyzed data from five waves of the European and World Values Surveys (WVS) (European Values Study Group and World Values Survey Association 2010) from 221 separate surveys conducted in 88 countries, with data from 317,109 individuals. Self-reported religiosity was measured along three dimensions: organizational participation, importance of God, and private religious identity. Well-being was operationalized by self-reported happiness and health. The results confirmed a positive link between self-reported religion and greater happiness and better self-reported health, but this was primarily in societies where religion is freely and widely practiced. In other words, “[t]he results confirmed a positive link between self-reported religion and greater happiness and better self-reported health, but this was primarily in societies where religion is freely and widely practiced. In other words, “the positive association between religion and well-being is not universal but depends upon the right to express religion freely and the opportunity to practice with like-minded others.” (Hayward and Elliott 2014, p. 23).

But the measurement of well-being can be operationalized differently. Instead of considering whether people report being happy, some researchers define well-being based on life satisfaction. Thus, in a recent meta-analysis, Yaden et al. (2022) investigated the relationship between religion/spirituality and life satisfaction. To this end, they conducted a meta-analysis of 256 articles, covering more than half a million participants. In this study, life satisfaction is defined as “a specific form of well-being”, “the overall assessment of how one’s life is going as a whole” (Yaden et al. 2022, p. 4150), it being understood that the subjective feeling of well-being is also, more broadly, associated with other aspects such as the absence of psychological distress or the feeling of momentary happiness. Only publications for which the effect size can be calculated were retained. Religiosity is measured either globally or by distinguishing, depending on the publication, one or more dimensions (religiosity, spirituality, religious attendance, religious practice, religious/spiritual experience). Depending on the dimensions considered, the effect size (calculated by r) varies between 0.11 (religious attendance) and 0.30 (spirituality), which the authors interpret as a medium effect based on Bosco et al. (2015). It is noted that the relationship between religiosity and life satisfaction becomes even more pronounced with age.

Another recent study adopts a many-analysts approach to measuring the relationship between religiosity and well-being (Hoogeveen et al. 2023). In this study, data from 10,535 individuals from 24 countries were analyzed by 120 different research teams. All 99 teams that used statistical analyses to calculate the effect concluded that it was positive, and 88% of them concluded that there was a significant link between religiosity and well-being. However, the median for the effect size is only ß = 0.12, which means that for almost half of the teams, the link between religiosity and well-being is negligible. For the other half, it remains small, except for two teams that found a medium effect. In any case, since the characteristics of the religious and spiritual profiles of the participants in these studies were not distinguished, the links between religion and well-being remain difficult to interpret. It cannot be excluded that the majority of participants who answered positively to the questions related to religion were precisely those for whom it was a source of encouragement. Thus, if those for whom religion was a disappointment or a burden simply preferred to report having no religion, the resulting data to be analyzed could have been skewed in favor of the expression of religion positively contributing to well-being.

Ultimately, it seems that most studies examining the relationship between religiosity and well-being struggle to provide very conclusive results. First, the vast majority of studies use a cross-sectional design, which does not allow for capturing the temporal order of the links between variables. Is it a change in religiosity that is followed by a change in the subjective assessment of well-being or vice versa? To be able to take into account the temporal order of subjective experiences and practices, longitudinal studies are necessary. However, most often, longitudinal studies do not highlight a significant relationship between religiosity and well-being. This is highlighted not only by the meta-analysis conducted by Yaden et al. (2022) on the relationship between religion/spirituality and life satisfaction, but also by the systematic reviews and meta-analyses conducted by Pankowski and Wytrychiewicz-Pankowska (2023a, 2023b) on the link between religious coping and mental health during the COVID pandemic.

Second, studies generally focus on statistical significance rather than effect size. However, as Prati (2024) shows, claiming that religion promotes well-being based solely on the statistical significance of a difference is a far too general a statement, difficult to interpret practically. To demonstrate this, he conducts two studies to measure the effect size of the measured links between religious indicators (affiliation, collective or individual practices, etc.) and well-being. The first of these studies uses merged data from two very large-scale surveys (European Values Study; World Values Survey) for a total of 645,249 participants from 115 countries. The second study uses data from three longitudinal surveys (the Wisconsin Longitudinal Study, the Survey of Health, Ageing, and Retirement in Europe, and the Midlife in the United States). In both cases, the estimated effect size for the links between religious indicators and well-being reported by participants are very small, if not negligible. In other words, these links are of little use in explaining what determines well-being compared, for example, to income, whose “effect size was approximately 150% larger than that (converted to Cohen’s d) for religious importance or religious participation (Study 1)” (Prati 2024, p. 374). There are several reasons why effect sizes are small and negligible when religiosity is measured in a non-differentiated manner. As Rosmarin and Koenig (2020) show, there are forms of religiosities that are soothing and others that are anxiety-provoking. For example, believing in a loving, caring, benevolent, kind, and generous God does not induce the same feelings as believing in a God who criticizes, judges, punishes, destroys, and is malevolent. Speaking of religion or spirituality in general is therefore too vague. Some forms of religiosities promote well-being, while others threaten it. The lack of a relationship between spirituality and quality of life in older people facing aging, as observed for example by Molzahn (2007), may simply result from the fact that different forms of spirituality have not been distinguished.

Third, well-being depends more on having a worldview that gives confidence than on religious beliefs per se (Galen and Kloet 2011). Some non-religious people who have a worldview that gives them confidence will report a greater sense of well-being than people with religious beliefs that leave them feeling uncertain.

Fourth, the definition of well-being varies across studies. In some cases, it is operationalized by measuring the feeling of happiness (Hayward and Elliott 2014), in other cases by life satisfaction (Yaden et al. 2022), by quality of life (Bartlett et al. 2003; Davison and Jhangri 2013; Molzahn 2007), self-reported health (Hayward and Elliott 2014), or other indicators such as the reduction in stress or anxiety in the face of illness, aging or death. Therefore, comparisons between studies that examine the relationship between religion and well-being must ensure the comparability of the indicators used to define well-being.

Finally, along the same lines, comparisons between studies must ensure the comparability of the indicators used to measure religion or spirituality. From this perspective, the longitudinal study conducted in Germany among 6653 adults over the age of 18 by Steinmann et al. (2024) is very informative. This is a cohort for which data have been available since 1984. Before the COVID-19 pandemic, a positive association was observed between religiosity (measured by attendance at religious services) and subjective well-being. The significant difference between religious and non-religious people lost its statistical significance during the first months of the COVID pandemic. A greater number of religious people reported a decrease in benefits in terms of well-being during this period. Analyses show that this decrease is explained by the decline in social integration of these individuals. This suggests that, if we define spirituality, for example, by religious affiliation, we must take into account the fact that it is a factor of social bonds and integration (Ugur and Aydın 2023; Ysseldyk et al. 2013), even if it is by far not the only means of socialization (Joiner et al. 2002).

In other words, to be able to interpret the links between religion and well-being, it is always important to specify the form of religiosity considered. This is well highlighted by the systematic literature review conducted by Litalien et al. (2022) to answer the question of whether religion and spirituality influence health behaviors and well-being. The review covers the period 2000-April 2019 in Canada. 128 publications were selected. Among the 38 dealing with mental health, several have results that present a U-shaped curve for the relationship between religion/spirituality and mental health (Badyal 2003; Balbuena et al. 2013; Blazer 2009; Dilmaghani 2018; Koenig 2009; Rasic et al. 2011; Rickhi et al. 2015; Schieman 2008; Snowshoe et al. 2017). Highly religious people and secularized people cope better with mental illness than people with average religiosity or without strong beliefs related to religion/spirituality.

Furthermore, religiosity is multidimensional (Brandt 2019). It includes cognitive (worldview), affective (bond of attachment to a spiritual figure), collective (membership in a community), ethical (promoted or prohibited behaviors), and identity (roles) aspects that are translated into beliefs and practices. Religious traditions provide resources to face life’s difficulties in all these dimensions. When comparing how various religious traditions contribute to well-being and quality of life, it is important to differentiate between the different dimensions listed.

2.2. Protestants, Catholics, and Well-Being

More than twenty years ago, in a study on the influence of religious integration on religious attitudes based on an international survey (ISSP Research Group 2000), Bréchon (2002) noted that in Europe, regardless of the country considered, integration into Protestantism or Catholicism resulted in more traditional and conformist attitudes, without clear differences between the two, stating that Catholics and Protestants are “culturally quite close today” (Bréchon 2002, p. 461). This did not prevent him, a few years later, from stating that “Religious practice has resisted better in Catholic countries than in Protestant ones” (Bréchon 2008, p. 27), because the traditional obligation to participate in Sunday celebrations is clearly stronger in the former than in the latter. In this sense, and more recently, he notes that the erosion of religious affiliation is much stronger in Protestantism than in other religious denominations. In Europe in 2017, the affirmation of non-affiliation is much more marked in historically Protestant countries than in Catholic and Orthodox countries, even if the disparities between countries are great (Bréchon 2021, p. 312). Thus, as an extension of the question of the link between religiosity and well-being, it is worth integrating the difference between religious affiliations into this question, especially in Western Europe between Protestants and Catholics.

The review presented in the previous section shows that the association between religion and well-being appears to be positive across different religious groups. At the same time, several studies indicate that levels of well-being can vary significantly among believers from different religious traditions. As Cohen and Johnson (2017) point out, each religious group has its own unique set of beliefs, social norms, ritual practices, values, and other characteristics that define it much like any cultural group. Consequently, religious groups may differ not only in their overall levels of well-being but also in the factors that contribute to well-being within each tradition. Studies that compare well-being, happiness and mental health among members of different religions have confirmed that such differences do exist. Among them, the research by Ngamaba and Soni (2018), which provides a large-scale comparative analysis of variations in happiness and life satisfaction across different religious groups, presents particularly interesting results for our work, as it distinguishes between Protestants and Catholics. Based on data from the World Values Survey (1981–2014), which includes over 330,000 participants across 100 countries, researchers sought to understand how religious affiliation influences subjective well-being (SWB) and happiness, and to what extent this relationship is moderated by contextual factors such as a country’s economic and cultural development. The results of this study indicate that individual religiosity plays a significant role in shaping subjective well-being, alongside other factors such as the country’s level of development, health status, household financial satisfaction, and freedom of choice. Moreover, the findings show that adherents of certain faith traditions tend to report higher levels of happiness and life satisfaction than others. Their results show that Protestants report the highest average levels of happiness (mean of 3.21 on a scale from 1 to 4), followed by Buddhists (3.17) and Roman Catholics (3.13). They are followed by Jews (3.06), Hindus (3.05), Muslims (3.03), non-believers (3.02), and finally Orthodox Christians (2.72). In terms of life satisfaction (measured on a scale from 1 to 10), Roman Catholics report the highest mean score (M = 7.12; SD = 2.31), closely followed by Protestants (M = 7.07; SD = 2.33). Other groups score slightly lower, with a gradual decrease down to Orthodox Christians, who report the lowest average level of life satisfaction (M = 5.43; SD = 2.49).

Considering that numerous factors, especially the country’s level of development, can influence well-being and life satisfaction, studies conducted within a single country are particularly relevant for better understanding the potential differences between Protestants and Catholics. Unfortunately, to our knowledge, research directly comparing these two denominations within the same national context remains rather scarce. Among the studies that explicitly distinguish between these two congregations, Lim’s (2015) work stands out as a particularly robust example. Based on data from the Gallup Daily Poll, this study draws on an exceptional sample of over 1.3 million American citizens. In this large-scale analysis, the author categorized respondents into eight distinct religious traditions: Protestants, Catholics, Mormons, Jews, Muslims, and other non-Christians. The findings first reveal that individuals who attend religious services regularly, regardless of their tradition, report significantly higher levels of life satisfaction and a more positive–negative affect balance compared to those who participate infrequently or not at all. Furthermore, when examining variations according to religious affiliation, significant differences were identified: overall, individuals from Christian traditions—particularly Protestants—show higher levels of life satisfaction than non-Christians. Regarding affect balance (measuring the predominance of positive over negative emotions), Protestants stand out once again, followed by Mormons and Catholics (Lim 2015).

Among the studies that aim to compare Protestants and Catholics within the same country, the Irish study is also significant due to the size and composition of its sample. Lewis et al. (2011) conducted their research using data from the Ireland Health and Social Wellbeing Survey (2001). This survey was based on a representative sample of 5205 adults aged over 16. The denominational breakdown of the sample was as follows: 1806 Catholics (40.4%) and 2513 Protestants (56.3%). It is important to note here, as in the studies cited above, that the “Protestant” category encompassed a broad range of denominations, including Presbyterians, the Church of Ireland, Methodists, Baptists, Free Presbyterians, Brethren, and unspecified Protestants. First, the results of this study highlighted a positive association between more frequent attendance at religious services and better psychological health. With regard to denominational differences specifically, the data revealed that Protestants attended religious services significantly less frequently than Catholics. Finally, in terms of psychological well-being, Catholics showed significantly higher scores on the General Health Questionnaire (GHQ), indicating lower levels of psychological well-being compared to Protestants. Another study worth mentioning is that of Tix and Frazier (2005), although their sample was much smaller and focused on a younger population (USA students). The authors aimed to examine whether religious tradition plays a moderating role in the relationship between intrinsic religiosity (that is, faith lived in a personal and internal way) and mental health, specifically anxiety and depression. The results showed that among Catholics, higher intrinsic religiosity was significantly associated with greater anxiety and depression, whereas among conservative Protestants, this religiosity was linked (albeit not significantly) to slightly lower levels of anxiety and depression. Among mainline Protestants, no clear association was observed. The authors suggest that the theological context and the practices specific to each tradition may influence the relationship between religiosity and mental health. For example, Catholicism’s emphasis on sin and reconciliation could potentially generate anxiety and guilt for some believers. Conversely, Protestant churches (particularly conservative ones) place greater emphasis on individual faith, grace, and a personal relationship with God, which may foster positive emotions and a motivation oriented towards optimistic perspectives (such as attaining heaven). The authors stress that these results demonstrate that religious traditions may have unique psychosocial advantages and drawbacks, and that it is important not to compare the overall value of religious traditions solely on the basis of mental health outcomes.

2.3. Research in Switzerland

Unlike the American and Irish contexts, which remain strongly shaped by religion, Switzerland follows a path of advanced secularization, similar to that observed in many Western European countries. Since the 1960s, census data have shown a steady increase in the number of individuals with no religious affiliation. Moreover, a significant portion of those affiliated with Switzerland’s two historical churches, Protestant and Catholic, now maintain a rather distanced relationship with their institutions.

Despite this trend toward secularization, some studies, though relatively few, have demonstrated a positive association between religiosity, well-being, and mental health. Monnot and Stolz (2016), using data from the 2012 Swiss Household Panel, found a statistically significant, albeit modest, correlation between religious practice and overall well-being. Their findings indicate that non-practicing individuals report an average satisfaction level of 7.6 (on a scale from 0 to 10), whereas weekly practitioners report a score of 8.15. Furthermore, individuals who identify as very spiritual or religious report a higher average satisfaction score (8.1) compared to those who identify as non-religious or non-spiritual (7.8). The authors emphasize that this positive effect of religiosity is mainly observed among those who see themselves as strongly spiritual or religious and who actively and regularly practice their faith.

In addition, several studies focusing specifically on older adults in Switzerland have highlighted the beneficial role of religiosity in health and well-being. For example, spiritual well-being has been identified as a protective factor against the desire for an accelerated death (Bernard et al. 2017), and a positive effect of religiosity on life expectancy was found in a longitudinal study (Lerch et al. 2010; Spini et al. 2001). Two qualitative studies conducted in long-term care facilities (EMS) in French-speaking Switzerland also demonstrated that spirituality serves as an important adaptive resource for coping with the challenges of aging and the proximity of death (Freudiger et al. 2007; Dandarova-Robert et al. 2016).

As for denominational differences, the study by Steiner et al. (2020) is among the few that have explored this aspect. Based on data from the 2007 Swiss Household Panel, the authors examined the impact of religious affiliation and religiosity on life satisfaction. They limited their sample to Swiss citizens over the age of 18 who identified as Christians (Protestant or Catholic) or as having no religious affiliation, in order to control for cultural bias and better isolate religious effects. Their regression analyses show that religious affiliation, whether Protestant or Catholic, is positively and significantly associated with subjective well-being. Individuals with a religious affiliation report higher levels of life satisfaction than those without one. Regular attendance at religious services also has a significant positive effect: attending at least once a week is associated with an increase of 0.59 points in happiness scores, a stronger effect than that of having a partner and cohabiting (+0.52 points). Regarding prayer, only daily practice is weakly positively correlated with well-being, while occasional or rare practice shows no significant effect. Finally, the study finds that Protestants report higher levels of life satisfaction than Catholics, a result that aligns with findings from other national contexts.

2.4. Protestants and Catholics in the Canton of Vaud

In Europe, following the Reformation, the principle of “cujus regio, ejus religio” led to a confessional division of the territory which confined Catholicism and Protestantism to relatively hermetic geographical spaces. In this context, as Bovay explains, “Switzerland has been a part of, along with Germany and the Netherlands, the trio of bi-confessional countries in Europe.” (Bovay 1997, p. 17, our translation). In the 19th century, Switzerland consisted of 22 cantons, all of which had state status within the Confederation. Consequently, religious affiliation was still one of the components of relations between the cantons at that time. However, as elsewhere in Europe and more broadly in the Western world, religious diversity has increased significantly since then, following the progressive recognition of religious freedom (freedom of conscience) at both federal and cantonal levels.

Thus, the canton of Vaud, which is the subject of our study, is traditionally a Protestant canton. In 1900, Protestants represented 86.3% of the population, Catholics 13.1%. We see that at that time, more than 99% of the population was affiliated with one of these two confessions. However, during the 20th century, the situation was reversed in a context that became more diverse. In 1970, Protestants (62.7%) and Catholics (33.3%) still represented more than 95% of the population, in a proportion of two-thirds/one-third. But between 1970 and today, things will accelerate. Between 2000 and 2010, Catholics (31.1%) become the majority group ahead of Protestants (28.9%). But together, these two confessions only represent 60% of the population because the group of “non-affiliated” (25.7%) has exploded and other Christian communities (5.6%) and Muslim communities (4.2%) have gained importance. Today, this trend has become even more pronounced. According to the 2023 figures, the “unaffiliated” have become the majority group (41.6%), while Catholics (25.2%) and Protestants (18.1%) have declined and other Christian communities (6.3%) and Muslim communities (5.9%) have increased slightly.

This religious context has the advantage of allowing the comparison of Protestants and Catholics living in relatively equal proportions in the same social and cultural environment for several decades. For the elderly population on which our study focuses, the data collected indicate that declared religious affiliation during life is approximately 80%. In other words, it will be possible to compare two groups that are largely in the majority compared to the reference population (people aged 65 and over) and observe whether religious affiliation influences the way people practice religion, react to life’s difficulties and find well-being. This comparison will not only focus on declared religious affiliation, but also on the fact that people identify with this affiliation, whether they say they are religious or not.

3. Research Questions and Hypotheses

The literature review on the links between religiosity/spirituality and well-being shows that the interpretation of the results depends on how religiosity/spirituality and well-being are measured. Therefore, in our research, different measures of religiosity/spirituality as well as well-being are compared.

3.1. Research Questions

Focusing on self-reported religiosity, we distinguish between religious affiliation, self-identification as religious, spiritual and/or atheist and the frequency of different individual or collective religious/spiritual practices: participation in religious services, prayer, discussion of religious/spiritual topics, following celebrations in the media (radio, TV), meditation, and reading religious/sacred texts. This translates into a first objective of examining whether the results vary according to the way in which religiosity/spirituality is defined: does the corroboration or refutation of a link between religiosity/spirituality and well-being depend on the way in which religiosity/spirituality is measured? And, since the declaration of religious affiliation specifies the religious tradition of belonging, it is also possible to compare the links between religiosity and well-being according to denomination. Hence the questions: Is there a difference between Catholics and Protestants in this regard? And does it make any difference whether we stick to affiliation alone or if we consider more specifically Catholics and Protestants who also identify as religious?

Similarly, for the measurement of well-being, we consider five indicators in order to compare how they are related to the various measures of religiosity/spirituality. These five indicators are: self-reported happiness, quality of life, life satisfaction, subjective assessment of health and subjective assessment of financial situation. A second objective of our study is to test whether the results vary depending on which indicators are considered to measure well-being.

Furthermore, since the study was conducted a few months after the semi-confinement period due to the COVID-19 pandemic, it is relevant to consider whether the findings regarding the links between measures of religiosity/spirituality and well-being are affected by this context. In a previously published paper, three preliminary analyses using multiple linear regression models were conducted only on the client sample (Dessart et al. 2024). The aim was to identify variables that contribute to anxiety-depression, happiness and life satisfaction. According to these analyses—among other factors such as meaning in life or quality of life—religious identity is only a predictor for life satisfaction. This confirms the value of further exploring the links between religiosity/spirituality and well-being.

Regarding the specific impact of the pandemic, the analyses indicated a marginally significant association between experiencing a sense of insufficiency in the religious/spiritual domain and feelings of greater anxiety and lower life satisfaction. In other words, it is important to know whether religious people coped better or worse with deprivation, whether in the religious/spiritual domain or in other domains, than non-religious people.

Given that the set of questions is already very broad, this study unfortunately does not allow for further comparisons between positive (e.g., comforting) and negative (e.g., guilt-inducing) religiosity.

3.2. Research Hypotheses

The research questions are concretized in the following hypotheses:

3.2.1. Religious Identity and Practice: Comparison Between Protestants and Catholics

Hypothesis 1:

The differences between Protestants and Catholics will be more pronounced among people who identify as religious compared to those who report an affiliation (currently) while not identifying as religious.

This first hypothesis is somewhat preliminary. It does not directly address the link between religiosity and well-being, but serves to ensure that when comparing links between various forms of religiosity and health, it is better to consider people who declare themselves as being religious than to simply rely on a declaration of affiliation.

3.2.2. Links Between Religiosity/Spirituality and Well-Being

Hypothesis 2:

There will be positive links between all indicators of religiosity (affiliation, self-declaration of religion, various religious practices) and all five indicators of well-being considered (happiness, quality of life, life satisfaction, self-reported health, self-reported financial situation).

This second hypothesis does not distinguish between Protestants and Catholics or between different ways of measuring religiosity, but provides a first step into examining whether the mere declaration of a form of religiousness is sufficient for measuring links between religiosity and well-being. While the literature struggles to determine which measure best operationalizes well-being, we address the strong assumption that the choice of measures of religiosity and well-being is indifferent: if a link is corroborated, it should be observable regardless of the specific measures chosen.

3.2.3. Links Between Religious Identification and Well-Being

Hypothesis 3:

All of the links mentioned in Hypothesis 2 will be stronger among people who declare themselves not only affiliated but also identify as religious.

This hypothesis is complementary to Hypothesis 2. It assumes that the links between measures of religiosity and well-being will be stronger when religiosity is measured by self-identification of religiousness combined with affiliation compared to when religiosity is only measured by self-declaration of affiliation.

3.2.4. Differences Between Protestants and Catholics in the Links Between Religion and Well-Being

Hypothesis 4:

Protestants will report a higher level of well-being than Catholics.

This hypothesis is based on studies that indicate that in different contexts, Protestants tend to show higher levels of life satisfaction and are more likely to declare themselves happy.

3.2.5. Impact of COVID

Hypothesis 5:

Religious people (based on affiliation or identity) coped better with deprivations outside of the religious domain than non-religious people.

This hypothesis addresses the claim that religion or spirituality helps individuals cope with difficulties, including losses or hardships.

Hypothesis 6:

Religious people (based on affiliation or identity) coped less well with deprivations in the religious domain than non-religious people. Following Hypothesis 1, the effect should be even more pronounced for people who declare themselves as being religious (identity) than those who only declare an affiliation. This hypothesis is complementary to Hypothesis 5: Although religious people may cope better with general difficulties than non-religious people (i.e., not affiliated nor identifying as religious), it is assumed they were nevertheless more negatively affected by a lack of offers in the religious domain than people for whom such offers were not expected.

Hypothesis 7:

People with a higher level of religious service participation coped less well with deprivations in the religious domain than those with a lower level of participation.

This hypothesis is complementary to Hypothesis 6: Since the semi-lockdown has mainly negatively impacted social contacts, it is assumed that religiously active people (i.e., practicing) mainly feel this deprivation in the religious domain with regard to collective practices.

4. Methodology

Our study adopted a mixed-method approach combining surveys and semi-structured interviews using a nested-sample design. This means that the people interviewed in the qualitative part are drawn from the larger quantitative sample that responded to the survey.

4.1. Populations and Questionnaires

Two populations were targeted: (1) elderly people (65+) engaged in organized volunteering (n = 1420); (2) elderly people (65+) receiving home care services (n = 3000).

To create the sample of senior volunteers, we contacted 62 associations and volunteer groups involved in volunteer work with seniors in the canton of Vaud. Of these, 42 associations and volunteer groups responded positively to our call for participation. This represents a sample of 1420 seniors engaged in social and charitable volunteering who received a paper questionnaire. The questionnaire also included an internet link (LimeSurvey). Thus, participants could respond either online or by mail. The questionnaire for senior volunteers is organized into four sections. The first section (questions 1 to 11) focuses on volunteer engagement. It distinguishes between formal volunteer engagement within an association or organization and informal engagement outside of an association or organization. It also asks about the types of associations in which volunteering takes place, the types of assistance or activities through which volunteering is expressed, the motivations for getting involved, and whether conversations with elderly beneficiaries lead to more personal themes or even address existential questions. The second section (questions 12 to 23) collects socio-demographic data: age, gender, level of education, professional activity, etc. The third section (questions 24 to 30) focuses on the person responding to the questionnaire’s perception of their quality of life and well-being, in general and during the pandemic. The fourth section (questions 31 to 38) asks about values, spirituality, and religion. It collects information about the respondent’s religious affiliation and the religious affiliation of their father and mother when they were 12 years old. It also asks about the respondent’s beliefs and practices.

To create a sample of home care service clients, we contacted the Vaud Association for Home Help and Care (AVASAD), which works with people aged 65+ living at home (N = 17,537). To constitute the representative sample of this population, AVASAD randomly selected 3000 clients to whom it sent our questionnaire by mail. The participants could respond either online or by mail. The questionnaire for the clients is divided into three sections. The first section (questions 1 to 19) focuses on the support received from the CMS and quality of life. In this regard, the person’s perception of their autonomy and mobility is examined, as well as whether they experience anxiety. The questionnaire asks about the activities they engage in and the importance they attach to the areas of friends, loved ones, health, physical activity, leisure, and religion or spirituality. The second section (questions 20 to 29) focuses on spirituality and religion. It also asks about the person’s beliefs and practices, whether it is important for them to be able to discuss spiritual or existential issues with someone, and if so, with whom. The third section (questions 30 to 42) collects sociodemographic data as well as the person’s religious affiliation, based on the same model as the questionnaire sent to senior volunteers.

As the data were collected during the COVID-19 pandemic, we also explored how participants experienced the restrictions associated with the semi-lockdown. Four questions, distributed across the first and last sections of the questionnaires, asked how the semi-lockdown (from March to June 2020) was experienced.

Of these two samples, 617 volunteers returned complete and valid questionnaires between August and December 2020, and 614 home care recipients between October 2020 and January 2021. Those who wished could indicate their consent to be contacted for an interview.

4.2. Interviews

After analyzing the religious and spiritual diversity of the respondents, we selected a select number of them to create as diverse a panel as possible. On this basis, semi-structured interviews were conducted with 43 volunteers and 35 home care clients to gain deeper insights into their experiences. Lengthening an average of one hour, they were conducted by LS and GD between January and August 2021. At the participants’ request, the interviews took place—with a few rare exceptions—in their homes.

4.3. Measures of Self-Reported Religiosity/Spirituality

4.3.1. Religious Affiliation

To measure religious affiliation, both questionnaires included a question, “Do you belong to a denomination or religious community?” to which the answer could be yes or no. If the answer was yes, the question was then asked: “If yes, to which denomination or religious community do you currently belong?” This was followed by a checklist: Roman Catholic, Reformed Protestant, Evangelical, Jewish, Islamic, or other (with a box to freely specify the affiliation). If the answer was no, the question was then asked: “If no: have you previously belonged to a denomination or religious community?” to which the answer could be yes or no. Afterwards, if the answer was positive, the same list of religious affiliations could be completed as those who currently had a religious affiliation.

4.3.2. Describing Oneself as Religious

To determine whether people considered themselves religious, both questionnaires included a question worded: “Would you say you are a religious person?” with four response levels: very, mostly, mostly not, or not at all. The combination of questions on religious affiliation and identity provides the opportunity to know whether people consider themselves religious Catholics or non-religious Catholics, religious Protestants or non-religious Protestants, etc.

4.3.3. Religious Practices

For religious practices, the question was asked: “During the 12 months preceding the semi-confinement (COVID), how often did you participate in a religious service or celebration, devote time to prayer, talk to other people about religious or spiritual matters, follow a religious or spiritual celebration on television, radio or on new media/specialized sites on the internet, devote time to meditation, read one or more religious books, such as the Bible, the Koran, the Torah or another sacred book. For each practice, seven levels of frequency were proposed: several times a day, every day or almost every day, at least once a week, at least once a month, between 6 and 11 times a year, between 1 and 5 times a year, never.

4.4. Measures of Subjective Well-Being

Several indicators were used, four of which were common to both questionnaires. Subjective assessment of quality of life was measured by the question “How would you describe or rate your quality of life?” The response allowed for a five-level assessment: very good, good, fairly good, not very good, not good at all. The same was true for the subjective assessment of general health and financial situation. For the subjective assessment of happiness, the question was worded “Right now, if you think about your life in general, would you say that overall you are…” with four response options: very happy, fairly happy, not very happy, not happy at all. For the fifth indicator, life satisfaction, its wording was slightly different in the two questionnaires. For CMS clients, the question was worded as follows: “All things considered, how satisfied are you with your life?” The response could be graded into seven levels: completely satisfied, very satisfied, fairly satisfied, neither satisfied nor dissatisfied, fairly dissatisfied, very dissatisfied, and completely dissatisfied. For the senior volunteers, the question focused on satisfaction with social relationships: “All things considered, to what extent are you satisfied with your relationships with those close to you?”, with the same seven response levels.

4.5. Measures of Lacks During the Pandemic

Both questionnaires asked for subjective assessments of perceived lacks in various domains, including religion/spirituality, during the period of semi-lockdown. Three insufficiencies in contacts and four insufficiencies in offer and assistance were mentioned. For insufficiencies in contacts, the question was formulated as follows: “If you think about the period of semi-confinement (COVID), did you feel a lack of contact with the following groups: friends, family and relatives, social contacts (outside of friends, family and relatives)?» For the insufficiencies in offer and assistance, the question was formulated as follows: “If you think about the period of semi-confinement (COVID), did you feel a lack of offer or assistance in the following areas: health, physical activity/gymnastics, entertainment/leisure, religion/spirituality?” For the three insufficiencies in contact and the four insufficiencies in offer and assistance, four possible answers were each proposed: yes strongly, yes a little, no not very much, no not at all. For the analyses of the links between religiosity/spirituality and insufficiencies, the insufficiencies were grouped into two categories: insufficiencies in the area of religion/spirituality on the one hand, and all other insufficiencies on the other.

4.6. Statistical Analyses

All data were analyzed using IBM SPSS Statistics, Version 29.0.2.0. The alpha level for all two-tailed tests was set at 0.05. The data analyses were conducted at two distinct levels: first, bivariate analyses to test hypotheses about the relationships between individual variables, followed by multivariate analyses to build explanatory models.

To test hypotheses regarding the relationships between religiosity and well-being, a series of non-parametric bivariate tests were conducted separately for the volunteer and client samples. These tests were chosen due to the ordinal and categorical nature of the variables and the non-normal distributions characteristic of the samples, which consisted of older adults (aged 65+) recruited from particular contexts (volunteering and home care clientship).

Spearman’s rank-order correlation (ρ) was used to measure the strength and direction of monotonic relationships for ordinal and dichotomous variables. The Chi-square test of independence (χ²) was used to compare frequency distributions between groups, with Cramer’s V calculated as a measure of effect size. In one specific case, a Chi-square Goodness-of-Fit test was also employed to compare an observed frequency distribution against a hypothesized theoretical distribution.

To examine the influence of religiosity as a predictor of well-being while controlling for multiple factors, the volunteer and client samples were combined. A series of binary logistic regression models was then constructed using this combined sample. A stepwise backward elimination procedure was employed to systematically reduce a full model to a final, parsimonious model containing only statistically significant predictors.

Finally, to explore a hypothesized indirect effect, a mediation analysis (Model 4) was conducted using the PROCESS v5.0 macro for SPSS (Hayes 2022). This analysis tested the extent to which a proposed mediator statistically accounted for the relationship between religiosity and well-being. The procedure utilized 5000 bootstrap samples to provide a robust estimate of the indirect effect.

5. Results

The presentation of the results begins with a description of the sociodemographic characteristics of the two samples considered in our study. Then, before presenting the links between religiosity/spirituality and well-being, it is necessary to understand the religiosity of the people studied. In terms of affiliation, two groups are mainly represented: Roman Catholics and Reformed Protestants. A comparison between these two subgroups is presented, both for affiliation, disaffiliation rates, self-declaration as religious and for the various forms of practices considered. Only then are the links between religiosity/spirituality and well-being presented and analyzed, before briefly discussing the impact of the COVID-19 pandemic. Readers primarily interested in the links between religiosity and well-being can skip Section 5.1, Section 5.2 and Section 5.3 and go directly to Section 5.4.

5.1. Sociodemographic Characteristics

The overall sociodemographic characteristics and the denominational affiliation of the two study samples are presented in

Table 1. Sociodemographic characteristics of the two study samples.

Sociodemographics	Volunteers (N = 617)	n (Total)	Clients (N = 614)	n (Total)
Sex (female)	65.8%	403 (612)	67.2%	405 (603)
Age (mean)	73.05 years	617	81.59 years	612
Education (highest achieved)
Mandatory school	6.1%	37 (609)	25.4%	155 (611)
Secondary II degree	52.4%	319 (609)	53.2%	325 (611)
Tertiary degree	34.8%	212 (609)	17.7%	108 (611)
Other	6.7%	41 (609)	3.8%	23 (611)
Marital status
Married	55.8%	344 (617)	33.3%	202 (606)
Widowed/Divorced	36.0%	222 (617)	56.3%	341 (606)
Single	7.6%	47 (617)	8.9%	54 (606)
Other	0.6%	4 (617)	1.5%	9 (606)
Retire (Yes)	97.4%	596 (612)	97.4%	591 (607)
Region of residence
Urban	65.8%	402 (611)	63.3%	373 (589)
Intermediate	19.6%	120 (611)	23.1%	136 (589)
Rural	14.6%	89 (611)	13.6%	80 (589)
Swiss nationality (Yes)	96.3%	594 (617)	89.8%	548 (610)

and

Table 2. Denominational affiliation of the two study samples.

Denominational Affiliation	Volunteers (N = 617)	n (Total)	Clients (N = 614)	n (Total)
Current Religious affiliation (Yes)	76.7%	465 (606)	66.1%	403 (610)
Roman Catholic	32.3%	150 (464)	36.6%	147 (402)
Reformed	58.4%	271 (464)	52.2%	210 (402)
Evangelical	5.2%	24 (464)	6.7%	27 (402)
Other Christian	1.3%	6 (464)	3.0%	12 (402)
Other Non-Christian	2.8%	13 (464)	1.5%	6 (402)
Current Religious affiliation (No)	23.3%	141 (606)	33.9%	207 (610)
Of which: Past Religious affiliation (Yes)	76.1%	105 (138)	56.8%	113 (199)
Roman Catholic	36.5%	38 (104)	34.5%	38 (110)
Reformed	55.8%	58 (104)	59.1%	65 (110)
Evangelical	3.8%	4 (104)	4.5%	5 (110)
Other Christian	1.0%	1 (104)	0.9%	1 (110)
Other Non-Christian	2.9%	3 (104)	0.9%	1 (110)

.

The average age is a little over eight years higher for CMS clients than for volunteers. This is not surprising, as volunteering requires sufficient health to be able to get involved. CMS clients, on the other hand, constitute a sample of people who need help and care at home, and therefore generally have poorer health, which deteriorates with age. The proportion of women is similar in both samples, i.e., 2/3 women to 1/3 men. This overrepresentation of women in the volunteer sample is due to the fact that there are more women than men who engage in social and charitable volunteering; in the sample of CMS clients, the overrepresentation of women is due to the higher life expectancy for women (Baeriswyl 2018; Lamprecht et al. 2020). We also note a higher average level of education in the volunteer sample than in the client sample, which is characteristic of volunteering compared to the general population (Baeriswyl 2018; Erlinghagen and Hank 2006). It is also not surprising that marital status is higher among volunteers than among clients: this is a consequence of the lower average age of volunteers. The distribution by region of residence is similar, indicating that the entire canton of Vaud is fully and equally represented in both samples. The very high proportion of people of Swiss nationality in both samples indicates that the elderly population is overwhelmingly Swiss. As for religious affiliation (current or past), it was significantly higher in the volunteer sample (94.1%) than in the client sample (84.6%), χ²(1, N = 1216) = 28.549, p < 0.001, V = 0.153. This result represents a small but statistically significant association.

As a whole, the population considered is characterized by a mainly Christian religious socialization. The elderly population of the canton of Vaud has primarily been socialized in Switzerland’s two historic churches: the Roman Catholic Church and the Reformed Protestant Church. Together, these two affiliations represent approximately 90% in each sample. Suffice it to say that for comparisons between different affiliations, only Catholics and Protestants are sufficiently numerous to allow for statistically valid tests. In both samples (Volunteers and Clients), the group of disaffiliated individuals was composed of significantly more former Protestants than former Catholics. Further details are presented in two tables with accompanying comments in Appendix A.1 and Appendix A.2.

5.2. Affiliation and Religious Identity

By combining the answers to the three questions “Would you say you are a religious person?”, “Would you say you are a spiritual person?”, and “Would you say you are an atheist?”, it is possible to distinguish various religious profiles. Theoretically, there should be eight, but since no one declared themselves to be both religious and atheist, the data collected allows us to distinguish six.

Table 3. Distribution of the six religiosity profiles among Catholics and Protestants for volunteers and clients.

	Volunteers
Profiles 1	Catholic (N = 150)	n (total)	Protestant (N = 271)	n (total)
RSnA	36.0%	54 (150)	25.5%	69 (271)
RnSnA	14.7%	22 (150)	15.5%	42 (271)
nRSA	0.7%	1 (150)	4.1%	11 (271)
nRSnA	23.3%	35 (150)	28.4%	77 (271)
nRnSA	10.7%	16 (150)	9.2%	25 (271)
nRnSnA	14.7%	22 (150)	17.3%	47 (271)
	CMS clients
Profiles	Catholic (N = 147)	n (total)	Protestant (N = 210)	n (total)
RSnA	34.7%	51 (147)	24.8%	52 (210)
RnSnA	18.4%	27 (147)	18.6%	39 (210)
nRSA	1.4%	2 (147)	0.5%	1 (210)
nRSnA	18.4%	27 (147)	19.5%	41 (210)
nRnSA	4.1%	6 (147)	6.2%	13 (210)
nRnSnA	23.1%	34 (147)	30.5%	64 (210)

¹ R/nR: positive/negative response to the question “Would you say you are a religious person?”; S/nS: positive/negative response to the question “Would you say you are a spiritual person?”; A/nA: positive/negative response to the question “Would you say you are an atheist?”.

shows the distribution among Catholics and Protestants for the two samples of volunteers and clients.

The proportion of people who answer positively to the question “Would you say you are a religious person”, whether they also consider themselves spiritual (RSnA profile) or not (RnSnA profile), is well above the 17.5% of Swiss who identify with religious institutions and corresponds to the institutional type of the study by Stolz et al. (2016). In the group of volunteers, 51% of Catholics say they are religious compared to 41% of Protestants, in the group of CMS clients 51.2% of Catholics compared to 43% of Protestants. This is explained by the average age of our samples. In Switzerland, older people consider themselves more religious than the general population. We also note a very low proportion of people who identify as both spiritual and atheist (nRSA profile). Atheists, corresponding to the secular type (Stolz et al. 2016), are more likely to declare themselves non-religious and non-spiritual (nRnSA profile). Other well-represented profiles are those who identify as spiritual but not religious (nRSnA) and those who do not identify with either of these labels (nRnSnA profile).

On the other hand, when comparing Catholics and Protestants among the people who answer positively to the question “Would you say you are a religious person”, we observe a greater proportion among Catholics than Protestants. Beyond the simple dichotomy between religious/non-religious which is not statistically significant, the comparison between Catholics and Protestants across levels of self-reported religiosity (the degree to which they consider themselves very, mostly, mostly not, or not at all religious), with respect to their religious affiliation, only presents a significant, albeit small, correlation for CMS clients (Affiliation x Religiosity is significant, ρ = −0.140, p = 0.012): Catholic CMS clients tend to report higher levels of religiosity compared to their Protestant counterparts. And no difference appears between Catholics and Protestants of the two groups when considering those who declare themselves spiritual or atheist. Furthermore, we observe that Catholics and Protestants differed significantly in their proportion of religious-spiritual people (volunteers, χ²(1, N = 421) = 5.186, p = 0.023, Φ = −0.111; CMS clients, χ²(1, N = 357) = 4.155, p = 0.042, V = −0.108), although the effect sizes were small. However, the groups did not differ significantly in their proportion of non-spiritual religious people.

Therefore, in analyzing the possible links between religiosity and well-being, it will be necessary to verify whether the way in which religion is operationalized changes anything. Does considering affiliation alone give different results compared to self-declaration of oneself as religious? Or does defining religiosity by certain practices rather than by others lead to different results regarding the links between religiosity and well-being?

5.3. Religious Practices: Protestants Compared to Catholics

What about when comparing different religious affiliations? A series of Chi-square tests was initially conducted to examine the relationship between all current Christian affiliations (Catholic, Protestant, and Other Christian) and the frequency of engagement (more than once a week, less than once a week, never) in six different religious or spiritual practices.

While these initial tests yielded statistically significant results, the significance was an artifact of the very small “Other Christian” subgroup (see Table 2). Including this small group obscured any meaningful comparison between the two main denominations. Therefore, to ensure statistical validity and focus the analysis on an informative contrast, all subsequent denominational comparisons were restricted to participants identifying as either Catholic or Reformed/Protestant.

Six different religious or spiritual practices among senior volunteers and senior home care clients have been considered: attendance at religious services, prayer, discussion of religious/spiritual topics, following religious media (celebrations on TV, radio, internet), meditation, and reading sacred texts. Three self-identifications were considered: whether one considers themselves “religious” or not, “spiritual” or not, and “atheist” or not.

Unsurprisingly, within the sample of Catholics and Protestants, there are consistent significant associations between religious identity and each of the practices. The relationship between religious identity and a composite score of religious practices (the average of responses about the frequency of each practice) was both positive and strong for the volunteer sample (ρ = 0.56, n = 392, p < 0.001) and client sample (ρ = 0.53, n = 299, p < 0.001), meaning that individuals who consider themselves ‘religious’ engage in religious practices much more frequently than their ‘not religious’ counterparts.

Similarly, there is a consistent relationship between spiritual identification and each practice. The association between the former and the composite score is positive for both samples, but moderate for volunteers (ρ = 0.48, n = 392, p < 0.001) and strong for clients (ρ = 0.52, n = 299, p < 0.001).

The same is true when comparing people who say they are both religious and spiritual to those who do not say they are religious and spiritual. There is a consistent, positive, association between this combined religious-spiritual identification and each practice. The association between this identification and the composite practice score is strong for both volunteers (ρ = 0.55, n = 392, p < 0.001) and clients (ρ = 0.50, n = 299, p < 0.001).

For atheist identification, instead, there are also consistent associations with each practice but in the opposite direction. The association between the former and the composite score is negative and moderate for both volunteers (ρ = −0.39, n = 392, p < 0.001) and clients (ρ = −0.32, n = 299, p < 0.001).

The same is true among people who have an undefined religious identity (neither religious, nor spiritual, nor atheist) compared to those who declare an identity (religious or spiritual or atheist). The association between the agnostic identification and practices is moderately negative for volunteers (ρ = −0.41, n = 392, p < 0.001) and strongly negative for clients (ρ = −0.51, n = 299, p < 0.001).

No clear pattern is found when comparing the “spiritual-non-religious” with others. In fact, there is no significant association between this identity and the composite practice score within either sample. Among volunteers, this group shows lower attendance at religious services (ρ = −0.21, n = 417, p < 0.001), prays less frequently (ρ = −0.11, n = 410, p = 0.020) and reads sacred texts less often (ρ = −0.14, n = 412, p = 0.005). Among clients, the only significant relationship concerned meditation, which they do more frequently (ρ = 0.16, n = 327, p = 0.005). Furthermore, all of these individual associations are weak.

If we focus on the comparison between Protestants and Catholics, things become more nuanced when we look not only at current affiliation but also at those who self-identify as religious, whether they also claim to be spiritual or not, or even just religious (and not spiritual). Details of the comparison of Protestant and Catholic practices according to current affiliation, according to self-identification as religious, and according to self-identification as religious but not spiritual, can be found in Appendix A.3, Appendix A.4 and Appendix A.5.

From there,

Table 4. Frequencies of Religious Practice by Affiliation and Religious Self-Identification in the Volunteer sample.

		Affiliation (All)		Affiliation (R)		Affiliation (RnS)
Practice	Frequency	Cath.	Prot.	Cath.	Prot.	Cath.	Prot.
Attendance at religious services	At least once a week	43 (28.9%)	43 (16.0%)	40 (52.6%)	36 (33.0%)
	Less than once a week	74 (49.7%)	171 (63.8%)	34 (44.7%)	67 (61.5%)
	Never	32 (21.5%)	54 (20.1%)	2 (2.6%)	6 (5.5%)
Prayer	At least once a week
	Less than once a week
	Never
Discussion of religious/spiritual topics	At least once a week
	Less than once a week
	Never
Following celebrations (media)	At least once a week	25 (17.0%)	24 (9.0%)
	Less than once a week	53 (36.1%)	123 (46.2%)
	Never	69 (46.9%)	119 (44.7%)
Meditation	At least once a week
	Less than once a week
	Never
Reading religious/sacred texts	At least once a week	19 (13.0%)	64 (24.1%)	16 (21.6%)	52 (47.3%)
	Less than once a week	50 (34.2%)	71 (26.7%)	36 (48.6%)	38 (34.5%)
	Never	77 (52.7%)	131 (49.2%)	22 (29.7%)	20 (18.2%)

Note: Data are presented as n (%). The columns represent different sub-samples of self-identification: (All) = The entire sample of Volunteer participants; (R) = Participants who self-identify as ‘Religious’; (RnS) = Participants who self-identify as ‘Religious but not Spiritual’. Empty cells indicate that the Chi-square test of independence between affiliation (Catholic vs. Protestant) and practice frequency was not statistically significant (p > 0.05) for that specific sub-sample.

and

Table 5. Frequencies of Religious Practice by Affiliation and Religious Self-Identification in the Clients sample.

		Affiliation (All)		Affiliation (R)		Affiliation (RnS)
Practice	Frequency	Cath.	Prot.	Cath.	Prot.	Cath.	Prot.
Attendance at religious services	At least once a week	38 (27.0%)	29 (14.8%)	34 (44.2%)	23 (27.1%)	10 (37.0%)	7 (20.0%)
	Less than once a week	37 (26.2%)	85 (43.4%)	20 (26.0%)	44 (51.8%)	6 (22.2%)	20 (57.1%)
	Never	66 (46.8%)	82 (41.8%)	23 (29.9%)	18 (21.2%)	11 (40.7%)	8 (22.9%)
Prayer	At least once a week	98 (69.5%)	104 (54.5%)
	Less than once a week	20 (14.2%)	41 (21.5%)
	Never	23 (16.3%)	46 (24.1%)
Discussion of religious/spiritual topics	At least once a week	20 (15.2%)	20 (10.4%)
	Less than once a week	47 (35.6%)	95 (49.5%)
	Never	65 (49.2%)	77 (40.1%)
Following celebrations (media)	At least once a week	53 (38.4%)	51 (25.4%)
	Less than once a week	47 (34.1%)	72 (35.8%)
	Never	38 (27.5%)	78 (38.8%)
Meditation	At least once a week
	Less than once a week
	Never
Reading religious/sacred texts	At least once a week	22 (15.9%)	45 (23.2%)	17 (23.0%)	35 (42.7%)	2 (8.3%)	11 (33.3%)
	Less than once a week	25 (18.1%)	58 (29.9%)	16 (21.6%)	29 (35.4%)	1 (4.2%)	13 (39.4%)
	Never	91 (65.9%)	91 (46.9%)	41 (55.4%)	18 (22.0%)	21 (87.5%)	9 (27.3%)

Note: Data are presented as n (%). The columns represent different sub-samples of self-identification: (All) = The entire sample of Client participants; (R) = Participants who self-identify as ‘Religious’; (RnS) = Participants who self-identify as ‘Religious but not Spiritual’. Empty cells indicate that the Chi-square test of independence between affiliation (Catholic vs. Protestant) and practice frequency was not statistically significant (p > 0.05) for that specific sub-sample.

illustrate the diminishing differences in religious engagement between Catholic and Protestant volunteers and clients when personal religious identification is considered. Although a pattern of diminishing denominational differences emerges for most practices when narrower profiles are considered—especially in the client sample—a couple of core behavioral differences, notably attendance at religious services and reading sacred texts, tend to persist.

Regardless, the overall pattern refutes Hypothesis 1, as the differences regarding religious practice between Protestants and Catholics were not more pronounced among affiliated and religious individuals compared to those who are affiliated but distanced. In other words, those who are distanced appear more identitarian in their practices, while deeper religious identification tends to blur denominational differences.

5.4. Links Between Religiosity/Spirituality and Well-Being

Here are presented the links between religiosity and well-being depending on whether religiosity is measured by religious affiliation, by declaring oneself religious, or by religious practices. For each link, the analyses consider a diversity of indicators to measure well-being. This section tests hypotheses 2 and 3.

5.4.1. Religiosity Measured by Religious Affiliation

If religiosity is measured by religious affiliation (yes/no), significant associations are found for the volunteer sample only with self-rated health; volunteers affiliated with a religious community reported being significantly healthier than their unaffiliated counterparts, ρ = 0.13, n = 604, p = 0.009.

Similarly, for the client sample, a significant association was found only with a composite score of health-related quality of life indicators (the sum of responses across the dimensions of mobility, autonomy, usual activities, pain/discomfort, and anxiety/depression), where a lower total score reflects better health status. Religiously affiliated clients reported significantly better overall health than their unaffiliated counterparts, ρ = −0.09, n = 563, p = 0.043.

5.4.2. Religiosity Measured by Religious Identification

If religiosity is measured by the degree to which one considers themselves a religious person (according to a four-level ordinal scale from “not at all religious” to “religious”), no significant relationships are found with any of the indicators for well-being for the volunteer sample. However, when the variable is dichotomized by categorizing volunteers into two groups—a ‘religious’ group and a ‘not religious’ group (the latter comprising those who report being either “rather not religious” or “not at all religious”)—a significant association appears for happiness: religious people are more likely to report being happy than their not religious counterparts, ρ = 0.08, n = 606, p = 0.049.

Contrariwise, for the client sample, significant associations are found for happiness, quality of life, and satisfaction with life. In particular, the more religious clients considered themselves, the more they reported being happy (ρ = 0.12, n = 600, p = 0.004), having a higher quality of life (ρ = 0.10, n = 579, p = 0.020), and being satisfied with their life (ρ = 0.11, n = 594, p = 0.007). When clients are categorized as either religious or not, religious clients were more likely to report being happy (ρ = 0.10, n = 600, p = 0.018) and satisfied with their life (ρ = 0.10, n = 594, p = 0.019).

5.4.3. Religiosity Measured by Combined Religious Affiliation and Religious Categorization

If religiosity is measured by comparing individuals who are both affiliated and self-identified as religious to those who are only affiliated, no significant associations are found with any of the indicators of well-being for the volunteer sample.

However, for the client sample, those who were both religious and affiliated to a religious community reported being happier (ρ = 0.17, n = 396, p = 0.001), having better quality of life (ρ = 0.15, n = 377, p = 0.005), better self-rated health (ρ = 0.13, n = 379, p = 0.013) and more satisfaction with life (ρ = 0.13, n = 387, p = 0.008).

5.4.4. Religiosity Measured by Religious Practices

If religiosity is measured by the composite score of religious practices, a significant association is only found for the volunteer sample, with self-rated health: volunteers who practice their religion more often were less likely to consider themselves healthy, ρ = −0.09, n = 566 p = 0.041.

Similarly, for the client sample, a significant positive relationship is found between religious practice and the composite score for health: clients who practice their religion more often were less likely to be healthy, ρ = 0.11, n = 488, p = 0.020. They are also more likely to consider themselves happy, ρ = 0.10, n = 514, p = 0.024.

If the individual religious practices are examined separately, significant negative associations appear, for the volunteer sample, for attending religious services (ρ = −0.11, n = 599, p = 0.009) following religious media (ρ = −0.14, n = 595, p < 0.001) and reading sacred texts (ρ = −0.08, n = 592, p = 0.046); volunteers who practice these activities more often also report being less healthy.

Instead, for the client sample, a number of significant positive associations with well-being could be observed:

• Attending religious services is positively correlated with happiness (ρ = 0.14, n = 577, p < 0.001), quality of life (ρ = 0.09, n = 555, p = 0.040), satisfaction with life (ρ = 0.13, n = 569, p = 0.002), and satisfaction with friendships (ρ = 0.10, n = 505, p = 0.027).
• Prayer is positively correlated with the composite health score (ρ = 0.13, n = 529, p = 0.004), indicating a link with poorer health outcomes. Additionally, it is negatively correlated with financial status (ρ = −0.09, n = 542, p = 0.030); clients who pray more consider themselves as having a worse financial situation.
• Talking about religion or spirituality is associated with greater happiness (ρ = 0.09, n = 557, p = 0.029), quality of life (ρ = 0.13, n = 541, p = 0.003), and satisfaction with friendships (ρ = 0.15, n = 490, p = 0.001).
• Following religious media is positively correlated with happiness (ρ = 0.08, n = 575, p = 0.045).
• Meditation is associated with a higher quality of life (ρ = 0.11, n = 539, p = 0.009) and greater satisfaction with friendships (ρ = 0.09, n = 483, p = 0.044).
• Reading sacred texts correlates positively with happiness (ρ = 0.12, n = 562, p = 0.004), quality of life (ρ = 0.10, n = 547, p = 0.021), and satisfaction with life (ρ = 0.11, n = 554, p = 0.009).

5.4.5. In Summary

The results show great variability: depending on how religiosity and well-being are measured, different and sometimes contradictory conclusions are reached. This is particularly evident when comparing the two groups. For the volunteer sample, significant associations were almost exclusively limited to measures of health, with only one exception for happiness. In contrast, for the client sample, the multiple indicators of religiosity correlated with a much broader array of outcomes, including happiness, both measures of health, quality of life, life satisfaction, and satisfaction with friends.

Moreover, where significant associations were found, the effect sizes are consistently small (ρ < 0.30). Therefore, Hypothesis 2, which predicted consistent positive links between all measures of religiosity and well-being, is not confirmed.

Furthermore, Hypothesis 3 receives mixed support, with the pattern of results differing between the two samples. For volunteers, affiliation was linked to self-rated health while the measure of personal religious identity was not, though a dichotomized measure of identity was linked to happiness. However, if volunteers who were both affiliated and identified as religious were compared to those who were simply affiliated, no group difference could be found.

For the client sample, however, the data supports Hypothesis 3. Measures of self-identification as religious, as well as the combination of affiliation and identification, were more consistently and strongly linked to multiple indicators of well-being than simple affiliation was.

5.5. Differences Between Protestants and Catholics in the Links Between Religion and Well-Being

At the bivariate level, when taking both samples together, religious affiliation was significantly correlated with both self-rated financial situation and happiness: Protestants reported a better financial situation (ρ = −0.09, n = 754, p = 0.017) and a higher level of happiness (ρ = −0.104, n = 767, p = 0.004) than Catholics. No significant bivariate associations were found between affiliation and either and quality of life or self-rated health.

Further analysis of the happiness variable suggests that the distinction between Protestants and Catholics, found at the bivariate level, is specifically driven by the “very happy” category. The relationship holds when comparing the “very happy” group to all others (ρ = −0.104, n = 767, p = 0.004) but disappears under any broader grouping of the positive categories, such as combining “happy” and “very happy,” or even including the borderline “rather not happy” category.

This specificity was further tested using logistic regression. First, a model was run to predict the likelihood of being “happy” in a broad sense (controlling for age, financial situation, and marital status) within the combined sample (n = 745). In this initial model, and contrary to the general prediction of Hypothesis 4, religious affiliation did not emerge as a significant predictor.

However, when a second model was specified to predict the likelihood of being “very happy”, affiliation did emerge as a contributing factor (B = −0.57, p = 0.004), in support of Hypothesis 4. As detailed in

Table 6. Logistic Regression Predicting the Likelihood of Being “Very Happy”.

Independent Variable	B	S.E.	Wald	df	p	Odds Ratio [95% CI]
Age	−0.03	0.01	7.53	1	0.006	0.97 [0.94, 0.99]
Financial Situation (Ref: Very Good)			49.10	4	<0.001
Not at All Good	−2.59	1.06	5.96	1	0.015	0.08 [0.01, 0.60]
Not Very Good	−3.14	0.75	17.41	1	<0.001	0.04 [0.01, 0.19]
Rather Good	−1.72	0.29	36.28	1	<0.001	0.18 [0.10, 0.31]
Good	−0.82	0.24	11.45	1	<0.001	0.44 [0.28, 0.71]
Religious Identity	0.56	0.19	8.83	1	0.003	1.74 [1.21, 2.52]
Religious Affiliation	−0.57	0.20	8.24	1	0.004	0.56 [0.38, 0.83]
Constant	2.40	0.94	6.55	1	0.011	11.07

Note: CI = Confidence Interval. Ref = Reference Category. Religious identity was coded 0 = Not Religious, 1 = Religious. Religious affiliation was coded 0 = Protestant, 1 = Catholic.

, the overall model was statistically significant (χ²(7) = 91.48, p < 0.001) and accounted for approximately 17.3% of the variance in being “very happy” (Nagelkerke R² = 0.173). Holding all other variables constant, Protestants had 77.6% higher odds of being in the “Very Happy” category than Catholics. Furthermore, religious identity was also a significant positive predictor; individuals who identified as religious had 74.4% higher odds of being “Very Happy” compared to those who did not.

Furthermore, a supplementary mediation analysis was conducted with both samples (n = 745) to test whether the subjective evaluation of financial situation explains the relationship between denominational affiliation and the likelihood of being “Very Happy,” while controlling for age and religious identity. For the purposes of this analysis, the 5-level ordinal mediator (financial situation) was treated as a continuous variable. The overall model was statistically significant, explaining approximately 17.1% of the variance in being “Very Happy” (Nagelkerke R² = 0.171).

The results support a partial mediation model. A significant indirect effect was found for religious affiliation on being “Very Happy” through financial situation, B = −0.15, 95% Bootstrapped CI [−0.28, −0.04]. The direct effect of religious affiliation on being “Very Happy” also remained significant (B = −0.58, p = 0.004). Both covariates remained significant predictors: age was negatively associated with being “Very Happy” (B = −0.03, p = 0.006), while religious identity was positively associated with the outcome (B = 0.56, p = 0.003).

Another mediational analysis was conducted with a subsample of religious participants (n = 371) to test whether the same indirect and direct effects are found specifically for religious Catholics and Protestants. The overall model was statistically significant, explaining 22.1% of the variance in the outcome (Nagelkerke R² = 0.221). In this sub-sample, the indirect effect of religious affiliation on being “Very Happy” through financial situation was not statistically significant, with a 95% bootstrapped confidence interval that contained zero (B = −0.20, 95% Bootstrapped CI [−0.43, 0.01]). However, the direct effect of religious affiliation remained significant (B = −0.91, p = 0.001), whereas age was not a significant factor.

These results suggest that while financial situation is a factor in the happiness gap at the higher end (that is, the likelihood of being “Very Happy”) between the broader group of Catholic and Protestant participants of our study, it does not explain the same difference between the religiously committed Catholics and Protestants.

5.6. Religiosity and Lacks Experienced During the COVID-19 Pandemic

This paragraph presents the results of testing Hypotheses 5—7.

5.6.1. Religiosity and Protection from Lacks

Contrary to Hypothesis 5, which predicted that religiosity would buffer against non-religious deprivations, there were generally no associations found between religious identity and different insufficiencies during the pandemic. The only exception was within the client sample, where a stronger degree of religious identification was weakly but positively correlated with a perceived insufficiency of friends (ρ = 0.09, n = 572, p = 0.042) and a lack of social contacts (ρ = 0.09, n = 561, p = 0.045): clients who consider themselves more religious perceived slightly greater insufficiency in these two social domains.

5.6.2. Religiosity and Vulnerability to Religious or Spiritual Lacks

Consistent with Hypothesis 6, which predicted that religious people felt more strongly a lack of religious support or offer, religiosity was moderately linked to a perceived insufficiency in offer or support in the religious or spiritual domain for both samples: individuals with a stronger religious identity reported a greater lack of religious or spiritual offer or support during the pandemic, both for the volunteer sample (ρ = 0.48, n = 566, p < 0.001) and the client sample (ρ = 0.36, n = 555, p < 0.001, respectively). Likewise, when using the dichotomized measure, ‘religious’ individuals reported a greater perceived lack of offer or support in the religious or spiritual domain than their ‘not religious’ counterparts in both the volunteer (ρ = 0.44, n = 566, p < 0.001) and the client (ρ = 0.33, n = 555, p < 0.001) samples.

5.6.3. Religious Practice and Vulnerability to Religious or Spiritual Lacks

Hypothesis 7, which asserted that individuals with higher levels of religious service participation would more strongly perceive insufficiencies in the religious domain, is corroborated: the frequency of religious attendance in the year preceding the pandemic was significantly and positively correlated with a perceived lack of religious or spiritual offer or support during the pandemic for both samples. The association was strong for the volunteer sample (ρ = 0.57, n = 558, p < 0.001) and moderate for the client sample (ρ = 0.37, n = 537, p < 0.001).

6. Discussion

The discussion will focus successively on the forms of religiosity present in our samples, on the links between religiosity and well-being, on the comparison between Protestants and Catholics in this regard, on the impact of the COVID-19 pandemic, before concluding on the methodological lessons that can be drawn from it.

6.1. Affiliation, Disaffiliation, Religious Identity

With regard to current affiliation, those who declare a religious identity represent at best half of those who have a religious affiliation. This is in line with a distancing and a loss of confidence in religious institutions (Bréchon 2021; Stolz et al. 2022). The comparison between Catholics and Protestants shows a stronger tendency towards disaffiliation among Protestants and a slight tendency among Catholics to identify as religious. As expected, the erosion of religious affiliation is greater among Protestants than among Catholics. It is striking that these developments are also noticeable among older people, who constitute the generations with the most pronounced religious identity following the strictly denominational religious education received by those born before the Second World War. Indeed, those born after the Second World War experienced the ecumenical opening in their youth, with the consequence of blurring the importance of the differences between Protestants and Catholics. The results show the effect of this shift in mentalities among those who were 75 years old and younger when the data was collected in 2020.

6.2. Practices According to Affiliation and Identity

Self-reported identity profiles—‘Religious’, ‘Religious-Spiritual’, ‘Weak/Undefined’, and ‘Atheist’—proved to be consistently associated with engagement in religious practices among Catholics and Protestants. An exception was the ‘Spiritual but not Religious’ profile, which demonstrated a more complex relationship.

When considering both denominations together, participants identifying as ‘Religious’ or ‘Religious-Spiritual’ reported the highest levels of overall religious practice. Conversely, those identifying as ‘Atheist’ or having a ‘Weak/Undefined’ identity reported the lowest levels of engagement. The ‘Spiritual but not Religious’ profile lacked a statistical relationship with the overall level of religious practice. However, this identity was associated with a distinct style of engagement that differed by sample. Among volunteers, ‘Spiritual but not Religious’ individuals engaged less in traditional practices like service attendance, prayer, and reading sacred texts. In contrast, among clients, they engaged more in meditation.

When focusing on a comparison between Catholics and Protestants, the breadth of denominational differences shrinks among the most religiously committed individuals. Among the general population of affiliated volunteers, Catholics reported significantly higher weekly participation in communal activities like service attendance and following religious media, while Protestants reported significantly higher engagement in reading sacred texts. However, when considering the subgroup of ‘Religious’ volunteers, the difference for following religious media disappeared, and no differences remained for the Religious but not Spiritual subgroup.

A similar pattern of diminishing denominational differences was also observed among clients, though it was less pronounced than with the volunteers. In the general affiliated group, Catholics and Protestants differed significantly across all practices except meditation. When narrowing the focus to the ‘Religious’ (defined broadly to include both Spiritual and not Spiritual) and the ‘Religious but not Spiritual’ client subgroups, most of these differences disappeared. However, the core differences for attendance at religious services and reading religious/sacred texts persisted even among the ‘Religious but not Spiritual’ clients.

These results show that, contrary to our Hypothesis 1, when religious identity is more affirmed, the difference between Protestants and Catholics tends to fade. This is in line with what Bréchon states that Catholics and Protestants are “culturally quite close today” (2002, p. 461), but this is more so for those who are engaged than for those who are distanced. Moreover, in this elderly population, differences in practices linked to religious socialization during childhood remain marked: Catholics go to mass more than Protestants to worship, but Protestants read the Bible more than Catholics. But we can think that this difference will itself disappear, among the most religious first. This is shown by the group of religious (but not spiritual) volunteers, whose average age is 73, compared to just over 81 for CMS clients. This last comparison should be taken with caution because it is based on modest figures, in the order of ten. This comparison is based on larger numbers of people who have the characteristics of being religious but open to also identifying as spiritual. Among people who identify as Protestant and Catholic, the differences in practices tend to blur: Protestants who identify as religious participate in worship, and Catholics who identify as religious and spiritual read the Bible.

In conclusion, when comparing the link between religion and well-being, we must consider not only religious affiliation, but also whether people identify as religious or not. Protestants and Catholics who say they are affiliated but not religious may be more different than people who say they are religious.

6.3. Links Between Religiosity and Well-Being

The results show that whether a link between religiosity and well-being is found depends on how religiosity is measured. As the title of this article suggests, religious affiliation is not enough. Indeed, with this measure of religiosity, we find significant links, for both volunteers and clients, only with self-rated health. In contrast, if religiosity is measured by considering oneself a religious person or not, a significant link appears for happiness in the sample of volunteers and for happiness, quality of life, and satisfaction with life in the client groups. If religiosity is measured through various religious practices, other links still appear, positive in the client sample and even negative in the volunteer sample.

As the literature review highlighted, it is not always easy to understand what lies behind religious affiliation or religious self-declaration. Religious affiliation can be an indicator of belonging to a socioeconomically advantaged group than average. For older adults, it can be an indicator of integration into a network. Yet, precisely, it has been shown that a comfortable financial situation (Prati 2024) or good social support (Joiner et al. 2002; Ugur and Aydın 2023; Ysseldyk et al. 2013) contribute to a sense of well-being.

Therefore, claims of links between religiosity and well-being should be handled with caution. Especially since, when statistically significant links can be found, the effect size is generally small or negligible.

6.4. Comparison Between Protestants and Catholics

The same caution should be exercised when drawing conclusions from a comparison between Protestants and Catholics. Indeed, our study confirms what several previous studies have shown, that Protestants tend to report being happier than Catholics. However, our study shows that this relationship is complex. Notably, while a bivariate link between denomination and happiness can be found when considering both samples together, this link disappears in more complex multivariate models predicting happiness. However, the denominational difference re-emerges when these same models are specific to predict the highest level of happiness (“very happy”).

Furthermore, for the general affiliated population, this specific gap in happiness (between those who consider themselves “very happy” and those who do not) appears to be partially explained by the fact that the Protestants in our samples also tend to have greater financial satisfaction, which mediates the link between religiosity as measured by denominational affiliation and the highest level of happiness. However, when we focus on the more religiously committed (those affiliated who also self-identify as religious), the indirect effect of financial satisfaction disappears, while the direct effect of denominational affiliation persists.

In other words, when religion is central to a person’s life, the source of the denominational happiness gap appears to shift away from material factors. The benefits derived from religion itself may compensate for material discomfort (as represented, for example, by financial dissatisfaction). Therefore, the simple contrast of religious affiliation is too crude when we seek to understand what a person derives from their religion to feel happy.

6.5. Impact of the COVID-19 Pandemic

Contrary to Hypothesis 5, people who self-identify as religious did not experience fewer gaps in relationships or opportunities in the areas of social contact, health, physical activity, or leisure. In other words, being religious did not compensate for the feeling of a gap in areas other than religious or spiritual. In contrast, in the religious or spiritual domain, lacks were felt much more strongly by religious than non-religious people, which confirms our hypothesis 6. In terms of problem-focused coping (Lazarus and Folkman 1984), being religious did not provide resources to replace lacks in other areas. In terms of emotion-focused coping, the data collected through the questionnaires do not provide information on whether religious people had more resources to emotionally regulate lacks. The interviews provide additional information showing that, in some cases, exchanges were adapted by using media that avoided physical contact or that religious resources were sufficiently internalized to manage the lack. For example, Dave, 76 years old, says: “Precisely, they canceled the meetings, so as not to put us in danger! But also, uh, other people! (…) Now, we do it by phone or by letter. That way, we try (…) to continue to keep our faith strong.” Or Adrien, 78: “Well, I stopped everything. (…). But, it didn’t bother me too much because I have all my baggage inside. I have the Scriptures. I have everything I need.” Sylvie, 87, explains how spiritual references help her give meaning to her life: “But it really had an impact (…) I understood it at a certain point, at the moment when I said ‘I can recharge my batteries myself!’. I understood even more than before that I need others. But this time, it was real life! And like Sheikh al-Alawi, the founder, said, ‘We are the humanity of a body, and we are all a cell, this body.’ So, if I am well, the body is well! If everyone strives to be well, in good health too, well, we are doing good for the body. So, that also gives meaning to life” As for Victor, 80, he explains that his spirituality had had an effect on the way he experienced the COVID period by offering him examples taken from the Christian tradition: “It certainly has an effect. Because that is very important. Yes. Yes, certainly. There are examples, from the popes, or from Jesus, yes. Certainly. There are corrections that I must undertake based on the examples they gave us.”

Hypothesis 7 is also confirmed. During the semi-lockdown, attendance at religious services (worship services, masses) was no longer possible. It is therefore not very surprising that the people who suffered the most from the lack of religious and spiritual offerings were those who engaged in the most collective practices. This lack was felt more strongly by volunteers than by clients. This is explained by the fact that volunteers are on average younger and healthier, which made them more mobile to regularly attend religious services. They were therefore more strongly affected by the semi-lockdown than CMS clients, some of whom were already unable to attend worship or mass before the semi-lockdown.