Religious Transitions Among Baby Boomers From Young Adulthood to Later Life: Associations with Psychological Well-Being Over 45 Years

Abstract

We used classification analysis to examine change in religiosity among baby boomers from young adulthood to early old age and how religiosity transition patterns are associated with psychological well-being in later life. In addition, we tested the gender difference in the above association. We applied latent class and latent transition analysis to 392 baby boomers who participated in the Longitudinal Study of Generations in Wave-1 (1971) and Wave-9 (2016). We identified three classes describing religiosity at each wave (strongly religious, doctrinally religious, and weakly religious), and considered five types of change or stability in religious class membership from Wave-1 to Wave-9. Multiple regression with gender interactions revealed that men who stayed strongly religious over the period reported better psychological well-being compared to men who declined in their religiosity; no such pattern was found for women. Our findings suggest that maintaining strong religiosity over the life course was beneficial for baby boom men in later life.

Keywords

baby boomer religiosity psychological well-being young adulthood older adulthood latent class analysis

Religiosity plays an important role in the daily lives of many older Americans and a substantial literature has documented the positive relationship between religiosity and well-being in later life (e.g., Bradshaw & Kent, 2018; Koenig et al., 1988). However, less is known about how religiosity changes are manifest over the adult life span and how those changes influence well-being outcomes from young adulthood to early old age. Further, much prior research on this topic has been limited by use of a specific dimension of religiosity (Hall et al., 2008) without taking into account its multidimensional nature (Ellison & Levin, 1998). To address these gaps in the literature, the current study took a person-centered approach—a statistical approach that identifies distinguishing patterns of characteristics in subgroups or classes of individuals (Scotto-Rosato & Baer, 2012)—to (1) identify religious classes among baby boomers in young adulthood and early old age, and (2) examine whether transitions across religious classes over time influenced psychological well-being over more than four decades of time.

The baby boom generation—consisting of individuals born between 1946 and 1964—occupies a unique position with regard to religious expression. This generation's exposure to social and political turbulence during their youth shaped an orientation to religion that was more individualized and less formally institutional than in previous generations (Roof & Mckinney, 1987). Scholars have noted that baby boomers were the first generation to view religion as a personal preference and to take a highly private approach toward religious matters ( Wuthnow, 1998). Although baby boomers were less religious than their parents’ generation in religious service attendance, confidence in religious institutions, and frequency of praying (Twenge et al., 2016), it is their engagement as religious “spiritual shoppers” in the religious marketplace that most differentiated them from their parents and grandparents.

The baby boom generation is an important cohort for research on religiosity and well-being, especially for understanding the search for meaning in this quantitatively and qualitatively unique cohort as it approaches late life (MacKinlay, 2014). Some evidence suggests that a substantial proportion of this generation has become more religious—with one study finding that over 20% perceived that they became more religious in their later years (Silverstein & Bengtson, 2018). Given the close links between religiosity, psychological well-being, and healthy aging, baby boomers provide a unique generational vantage point for understanding how religious change and stability over the adult life span influence psychological well-being in later adulthood.

Literature Review

Multidimensional Construct of Religiosity

Religiosity has long been considered a multidimensional construct (Cornwall et al., 1986; Verbit, 1970) consisting of “behavioral, social, doctrinal and denominational characteristics” (Fetzer Institute, 1999). Yet, there is little consensus regarding the key conceptual domains and measurements of religiosity. Studies of religiosity and well-being outcomes have tended to rely on a single item to measure religiosity (most often religious service attendance, Hall et al., 2008), that neglect the complexity of the construct (Ellison & Levin, 1998). Further, multidimensional treatments of religiosity often use summative scores that combine conceptually disparate dimensions (Allport & Ross, 1967; Hall et al., 2008), thereby ignoring heterogeneity in the patterns of those dimensions. In place of a variable-centered approach (i.e., factor analysis), which is limited by its assumption of linear additivity of measures, a person-centered approach (i.e., latent class analysis) provides distinct advantages in identifying clusters or classes of individuals that share observed characteristics (Collins & Lanza, 2010). A person-centered approach identifies subgroups that probabilistically share observed characteristics by optimizing within group homogeneity and between group heterogeneity (Scotto-Rosato & Baer, 2012). For this reason, person-centered approaches have been useful in typological studies.

In the study of religiosity, researchers have used person-centered approaches (i.e., latent class analysis) to examine patterns of religious expression and identify typologies or profiles that capture the multidimensional and heterogeneous nature of religiosity (Klemmack et al., 2007; Park et al., 2013). For example, Park et al. (2013) identified four classes of religiosity among adults in the United States based on religious service attendance, prayer, positive religious coping, and daily spiritual experiences. Similarly, Klemmack et al. (2007) identified six types of religiosity among older adults in the United States using three indicators: organizational involvement, nonorganizational involvement, and intrinsic religiousness. However, little research has been conducted classifying the religiosity of baby boomers across disparate stages of their life course.

Religiosity and Psychological Well-Being

A considerable amount of literature has investigated the relationship between religiosity and psychological well-being (for reviews, see Hackney & Sanders, 2003; Koenig & Larson, 1998; Koenig et al., 2001; Levin & Chatters, 1998). Psychological well-being refers to an individual's subjective evaluation of the quality of their life, consisting of two components—cognitive and affective (Diener, 1984; Diener et al., 2018). Koenig et al. (2001) found in their extensive and systematic review of the literature on religion and psychological well-being that 80% of the reviewed studies found at least one significant positive correlation between religiosity and well-being. In addition, using pooled data from the General Social Surveys between 1972 and 1982, St. George and McNamara (1984) found that, regardless of race and gender, strength of religious identification was significantly related to global happiness among middle-aged adults.

Mixed findings have also been reported from research on religiosity and psychological well-being. For example, Yoon and Lee ( 2004) reported a positive relationship between religiosity and the cognitive dimension of well-being (i.e., life satisfaction) among rural African American and Native American older adults. However, other research failed to find a significant relationship between religious belief and emotional well-being (Mak et al., 2011). Further, Leondari and Gialamas (2009) found that church attendance, but not other aspects of religion, was positively associated with life satisfaction in a sample of young adults. Inconsistent findings were also found in the linkages between religiosity and the affective dimension of well-being. For example, Zautra et al. (1977) found no relationship between religious participation and happiness, while two other studies found weak correlations between religiosity and positive affect (Fry, 2000; Lun & Bond, 2013).

Studies using person-centered approaches have generally found evidence of a positive relationship between religiosity class membership and well-being. For example, Bravo et al. (2016) found that college students in low and questioning religiosity classes reported lower levels of psychological well-being, and using a nationally representative survey, Park et al. (2013) found that individuals in a highly religious class reported fewer depressive symptoms than those in a low religious class.

Gender is a key feature when considering religious involvement and its benefits. Although, in general, women are more involved than men with religion (Leondari & Gialamas, 2009; Miller & Stark, 2002), in some studies men appear to have received greater rewards than women from religious attendance and intensity in the form of better subjective assessments of health (Mukerjee & Venugopal, 2018) and lower psychological distress and greater happiness (Maselko & Kubzansky, 2006). These differences are typically explained by the social integration provided men by religious communities, as well as the higher status accorded men in some religious denominations. However, other studies have found that religiosity is more strongly associated with lower risk of depression for women than for men (Strawbridge et al., 2001). Thus, findings with respect to gender differences in the positive impact of religiosity on emotional well-being are equivocal, as is likely the case with respect to this issue later life, about which even less is known.

The Current Study

This investigation takes a multidimensional person-centered approach to religiosity in order to identify clusters of individuals by their patterns of religious attributes in young adulthood and early old age, track translations in cluster membership over time, and examine the impact of religious transitions in cluster membership on psychological well-being in later life. We focus on religious attributes that include religious attendance, religious intensity, literal beliefs, and civic value of religion. We also investigate gender differences in the association between stability and change in religiosity and psychological well-being.

Method

Sample

This study used data from the Longitudinal Study of Generations (LSOG), a multigenerational and multitime point survey of 418 four-generation families with 3,686 respondents. The LSOG began in 1971 as a cross-sectional study with three-generation families, consisting of grandparents (G1), parents (G2), and grandchildren (G3) in the Southern California area. The G3s, who were 16−26 years old in 1971, represent the baby boom generation. In 1985, the LSOG became a longitudinal study and continued in 1988, 1991, 1994, 1997, 2000, 2005, and 2016. This study uses data collected from the G3 generation in 1971 and 2016. To collect data, mailed-back questionnaires were used in 1971 and both online and mail-back questionnaires were used in 2016 (for more details, see Silverstein and Bengtson, 2019). Our analytic sample contains 392 G3s in the early to middle part of the baby boom generation who participated at Wave-1 and Wave-9 and answered religiosity items in both waves.

Measures

Religiosity. Five measures of religiosity in Wave-1 and Wave-9 served as indicators in the latent class model. Religious service attendance was measured by one question: “How often do you attend church or religious services these days?” Response options ranged from (1) never to (6) more than once a week, and these six responses were dichotomized into high attendance (at least once per month) and low attendance (less than once per month). Religious intensity was measured by one question: “Regardless of whether you actually attend religious services, do you consider yourselves to be religious?” Response options ranged from (1) not at all religious to (4) very religious; these four responses were dichotomized into high religious intensity (very religious/moderately religious) and low religious intensity (somewhat religious/not at all religious). Civic value of religion was measured by two questions: “This country would be better off if religion had a greater influence on daily life” and “Every child should have religious instruction.” Response options ranged from (1) strongly disagree to (4) strongly agree, and four responses were dichotomized into high value (strongly agree/agree) and low value (disagree/strongly disagree). Finally, literal beliefs were measured by one question: “God exists in the form as described in the bible.” Response options ranged from (1) strongly disagree to (4) strongly agree, and four responses were dichotomized into strong belief (strongly agree/agree) and weak belief (disagree/strongly disagree).

Psychological well-being. We assessed participants’ psychological well-being in Wave-1 and Wave-9 using ten items of the Bradburn Scale of Psychological Well-Being (Bradburn, 1969). This scale contains five questions assessing positive affect (e.g. “During the past few weeks, did you feel particularly excited or interested in something?”) and five questions assessing negative affect (e.g., “During the past few weeks, did you feel very lonely or remote from people?”). Response options were dichotomous: (0) no or (1) yes. After reverse coding the negative affect items, the ten items were summed to create an index ranging from 0 to 10, with higher scores indicating better psychological well-being.

Control variables. The following participant characteristics were controlled in our multivariate analysis: age, gender (0 = men, 1 = women), race (0 = others, 1 = White), education (from 1 = 8th grade or less to 8 = post-graduate degree), annual household income (from 1 = less than $10,000 to 21 = $200,000 or more), religious affiliation (0 = none, 1 = affiliated), marital status (0 = unmarried/unpartnered, 1 = married/cohabitate); retirement status (0 = no, 1 = yes). Age, gender, race, and religious affiliation were derived from Wave-1 and marital status, education, annual household income, and retirement status were derived from Wave-9.

Analytic Strategy

We used Latent Gold 5.1 (Vermunt & Madigson, 2016) to conduct a latent class analysis using five dichotomized religiosity indicators in Waves 1 and 9. Latent class analysis is a person-centered approach to identify unobserved subgroups (latent classes) within a population based on individual participants’ responses to a set of indicators (Nylund-Gibson & Choi, 2018). Based on this analysis, we identified religious classes among baby boomers in young adulthood and early old age. We used Bayesian Information Criterion (BIC), Akaike Information Criterion (AIC), and Sample-Adjusted Bayesian Information Criterion (SABIC) as criteria to select the optimal number of latent classes. When the optimal number of latent classes was the same in Waves 1 and 9, we conducted a latent transition analysis using a latent Markov model (Vermunt & Madigson, 2016). Latent transition analysis enables estimation of the transition probability in latent class membership from one time to the next (Collins & Lanza, 2010). In this analysis, participants were assigned a probability of transitioning religious class membership between Wave-1 and Wave-9. Lastly, we conducted the three-step Bolck–Croon–Hagenaars approach (Bolck et al., 2004) to examine what classes of transition in religiosity from Wave-1 to Wave-9 were associated with psychological well-being in Wave-9. This method enabled us to adjust classification errors to reduce bias when compared to the classify–analyze approaches (Bakk & Vermunt, 2015). The predictive models were run simultaneously with the latent class models, so that the effects of latent transitions are weighted by probabilities of transition type between Wave-1 and Wave-9. Full-information maximum likelihood estimation was used to account for missing values (Vermunt & Magidson, 2016).

Results

Sample Characteristics

Characteristics of the sample and the distribution of religion and psychological well-being variables are presented in Table 1. The mean age of the sample in Waves 1 and 9 was 19 and 64 years old, and 59.2% of participants were women. Most participants were White (92.1%) and mean annual household income ranged between $90,000 and $100,000 at Wave-9. Most participants (83.2%) reported affiliation with a religion in Wave-1, but affiliation decreased to 52.3% in Wave-9. Lastly, participants reported relatively higher levels of psychological well-being in Wave-9 than in Wave-1.

Table 1.
Descriptive Results Among Demographic and Study Variables.

Baby boomer (n = 392)

Wave-1 Wave-9

Variables Range n (%) M (SD) n (%) M (SD)

Age 19.29 (2.66) 64.29 (2.66)

Gender

Men 160 (40.8)

Women 232 (59.2)

Race

White 361 (92.1)

Black 4 (1.0)

Hispanic 14 (3.6)

Others 12 (3.1)

Education 1–8 5.61 (1.39)

Marital status

Married 270 (68.9)

Living with a partner 19 (4.9)

Separated/Divorced/Widowed 87 (12.5)

Single/Never married 16 (4.1)

Annual household income 1–21 10.35 (6.01)

Religious affiliation

Mainline protestant 100 (25.5) 35 (8.9)

Evangelical protestant 69 (17.6) 64 (16.3)

Catholic 79 (20.2) 37 (9.4)

Jewish 50 (12.8) 43 (11.0)

Mormon 15 (3.8) 14 (3.6)

Others 13 (3.3) 12 (3.1)

None 62 (15.8) 182 (46.4)

Retirement

Yes 177 (45.2)

No 215 (54.8)

Religiosity indicators

Religious attendance: high 142 (36.2) 107 (27.3)

Religious attendance: low 246 (62.8) 284 (72.4)

Religious intensity: high 240 (61.2) 151 (38.5)

Religious intensity: low 146 (37.2) 238 (60.7)

Religion had a greater influence: high 249 (63.5) 204 (52.0)

Religion had a greater influence: low 140 (35.7) 185 (47.2)

Children get religious instruction: high 300 (76.5) 232 (59.2)

Children get religious instruction: low 88 (22.4) 155 (39.5)

God exists as described in the Bible: high 203 (51.8) 182 (46.4)

God exists as described in the Bible: low 171 (43.6) 203 (51.8)

Psychological well-being 0–10 5.87 (1.80) 7.46 (2.02)

Religious Classes and Transitions of Class Membership From Young Adulthood to Later Life

After conducting the latent class analysis, the BIC, AIC, and SABIC scores showed that a three-class model was the best-fitting model for participants in both Waves 1 and 9 (see Table 2). Item response and latent class probabilities of these three religious classes are presented in Table 3. Using 0.5 as a cut-off to define classes from item response probabilities, we label the three classes as follows: (1) strongly religious (item response probabilities in all indicators were over 0.5), (2) doctrinally religious (item response probabilities in literal beliefs and civic value of religion were over 0.5 but religious attendance and religious intensity were under 0.5), and (3) weakly religious (item response probabilities in all indicators were under 0.5). Latent class probabilities showed that the strongly religious class was most common at Wave-1 (40%); however, in Wave-9, the weakly religious class was most common (40%).

Table 2.
Latent Class Analysis Statistics and Fit Indices.

Wave-1 Wave- 9

Model Classes (n) BIC AIC SABIC BIC AIC SABIC

Model 1 1 2491.01 2471.15 2475.14 2600.41 2580.56 2584.55

Model 2 2 2163.04 2119.36 2128.14 1986.07 1942.38 1951.16

Model 3 3 2171.13 2103.62 2117.19 1974.24 1906.73 1920.30

Model 4 4 2202.10 2110.76 2129.12 1998.68 1907.34 1925.70

Model 5 5 2232.21 2117.04 2140.19 2029.55 1914.38 1937.53

Model 6 6 2491.01 2471.15 2475.14 2061.64 1922.64 1950.58

Note. Bolded values indicate best fit for each respective statistic. BIC = Bayesian Information Criterion; AIC = Akaike Information Criterion; SABIC = Sample-size Adjusted Bayesian Information Criterion.

Table 3.
Item Response and Latent Class Probabilities of Three Religiosity Classes.

Item response probabilities Strongly religious Doctrinally religious Weakly religious

Religious attendance 0.75 0.16 0.05

Religious intensity 0.99 0.49 0.26

Wave-1 Religion had a greater influence 0.93 0.69 0.22

Children get religious instruction 0.99 0.91 0.36

God exists as described in the Bible 0.80 0.70 0.07

Religious attendance 0.73 0.09 0.03

Religious intensity 0.93 0.29 0.01

Wave-9 Religion had a greater influence 0.97 0.71 0.03

Children get religious instruction 0.99 0.75 0.18

God exists as described in the Bible 0.94 0.54 0.05

Latent class probabilities Strongly religious Doctrinally religious Weakly religious

Wave-1 0.40 0.27 0.33

Wave-9 0.32 0.28 0.40

Note: Bold indicates item response probabilities over 0.5.

Regarding the result of latent transition analysis (see Table 4), latent class transition probabilities from Wave-1 to Wave-9 showed that 76% and 55% of participants remained in weakly and strongly religious classes, respectively, from young adulthood to later life. However, only 37% of participants stayed in the doctrinally religious class from young adulthood to early old age and 51% moved from doctrinally to weakly religious classes over time.

Table 4.
Latent Transition Probabilities of Three Religiosity Classes.

Wave-1\Wave-9 Strongly religious Doctrinally religious Weakly religious

Strongly religious 0.55 0.27 0.18

Doctrinally religious 0.12 0.37 0.51

Weakly religious 0.13 0.11 0.76

Associations Between Religious Transitions and Psychological Well-Being

Latent transition analysis among the three classes identified yielded nine transition patterns that we reduced to five patterns in order to maintain adequate cell sizes and ensure adequate power for the subsequent analysis. These patterns are: (1) staying strongly religious (22.9%), (2) staying doctrinally religious (13.8%), (3) staying weakly religious (16.3%), (4) increasing religiosity by moving from doctrinally religious to strongly religious, weakly religious to strongly religious, or weakly religious to doctrinally religious (9.5%), and (5) decreasing religiosity by moving from strongly religious to doctrinally religious, strongly religious to weakly religious, or doctrinally religious to weakly religious (37.4%).

We conducted multivariate regression analyses to identify how membership in the five latent class transition patterns was associated with psychological well-being at Wave-9. In Model 1, we regressed psychological well-being on these religious latent class transition patterns, controlling for demographic variables, as well as baseline (Wave-1) psychological well-being to control for psychological predispositions in young adulthood and minimize the possibility that emotionally positive individuals are more likely to have or develop strong religious orientations. In Model 2, interaction terms between religious transitions and gender were added (Table 5).

Table 5.
Associations of Class Memberships with Outcome Variables (Reference: Staying Strongly Religious^a).

Psychological well-being Wave-9

Model 1 Model 2

Variables b (SE) z b (SE) z

Intercept 4.86 (1.11) 4.35* 5.75 (1.19) 4.79*

Class memberships

Staying doctrinally religious^b 0.05 (0.47) 0.10 −0.49 (0.77) −0.64

Staying weakly religious^c 0.02 (0.46) 0.05 −0.40 (0.53) −0.75

Increasing religiosity^d −0.44 (0.54) −0.80 −0.87 (0.63) −1.36

Decreasing religiosity^e −0.21 (0.32) −0.64 −1.18 (0.47) −2.51*

Covariates

Psychological well-being (Wave-1) 0.14 (0.06) 2.35* 0.15 (0.06) 2.58**

Age (Wave-1) 0.02 (0.03) 0.57 0.01 (0.03) 0.40

Women (Wave-1) −0.15 (0.19) −0.80 −1.03 (0.44) −2.33*

White (Wave-1) 0.16 (0.39) 0.41 0.20 (0.39) 0.51

Religious affiliation (Wave-1) 0.02 (0.32) 0.07 −0.05 (0.32) −0.16

Education (Wave-9) 0.05 (0.09) 0.62 0.03 (0.09) 0.34

Annual income (Wave-9) 0.04 (0.02) 2.05* 0.04 (0.02) 2.04*

Married/cohabitate (Wave-9) 0.69 (0.29) 2.37* 0.60 (0.29) 2.04*

Retired (Wave-9) 0.23 (0.20) 1.15 0.19 (0.20) 0.96

Interaction terms

Staying doctrinally religious × Women - - 0.77 (0.93) 0.82

Staying weakly religious × Women - - 0.53 (0.71) 0.74

Increasing religiosity × Women - - 0.58 (0.97) 0.60

Decreasing religiosity × Women - - 1.56 (0.62) 2.24*

^a
Participants who were in the strongly religious class in Waves 1 and 9 (22.9%).

^b
Participants who were in the doctrinally religious class in Waves 1 and 9 (13.8%).

^c
Participants who were in the weakly religious class in Waves 1 and 9 (16.3%).

^d
Participants who moved from the doctrinally religious to strongly religious, weakly religious to doctrinally religious, and weakly religious to strongly religious across Waves 1 to 9 (9.5%).

^e
Participants who moved from the strongly religious to doctrinally religious, strongly religious to weakly religious, and doctrinally religious to weakly religious across Waves 1 to 9 (37.5%).

p<.05. p<.01. p<.001.

Coefficients shown in Model 1 reveal no significant differences in psychological well-being between staying strongly religious and the other transition patterns. In Model 2, we tested for interactions between religious transition patterns and gender on Wave-9 psychological well-being. Results showed a significant interaction effect between gender and the decreasing religiosity pattern and gender (reference group: staying strongly religious). Figure 1 displays predicted values of psychological well-being of the four groups formed by the interaction.

Figure 1.
Predicted psychological well-being by gender and religious transition patterns (model 2 in Table 5).

In the staying strongly religious pattern, men reported higher levels of psychological well-being than women. In the decreasing religiosity pattern, however, women reported slightly higher levels of psychological well-being than men. In other words, men had significantly better psychological well-being in the staying strongly religious pattern compared to the decreasing religiosity pattern, whereas women's psychological well-being did not differ between the staying religious and decreasing religiosity pattern. Other interaction effects were not significant.

Discussion

This investigation aimed to provide a multidimensional person-centered depiction of religiosity in a sample of baby boomers measured in from young adulthood and early old age and identify associations between religious change and psychological well-being in later life. We identified the same three religious classes at both stages of life, and a fair degree of transitioning across types over the 45 year interval of the study. Over that interval, the class characterized as strongly religious declined from 40% to 32%, weakly religious increased from 33% to 40%, and doctrinally religious showed little change from 27% to 28%. At the individual level latent transition analysis showed that approximately 45% of those classified in the strongly religious type moved to doctrinally religious (27%) and weakly religious classes (18%) in later life. Three-quarters (76%) who were classified in the weakly religious type in young adulthood remained weakly religious in in their later years, the most stable religious class. Thus, our findings followed trends identified in national data of general religious weakening in the population over recent decades. However, our results should be uniquely interpreted in a life course context as distinct from national data which are not longitudinal in design. Thus, our findings indicate that baby boomers are not experiencing elevated religiosity in later life, as some commentators maintain; instead, our findings reveal a trend that is more consistent with historical trends toward secularism (Voas & Chaves, 2016). Developmental and cohort-specific processes may also contribute to these results as baby boomers were part of a cohort that achieved higher education in substantial numbers, a status related to secularization (Idler, 2021), and loosened their connection to the religion of their upbringing in midlife (Roof, 2001).

A person-centered approach also allowed identification of doctrinally religious class in both early and later adulthood as a mixed type of religiosity. This class is similar, though not identical, to the religious type described in the literature as privately religious (Suitor et al., 2018; Vasilenko & Espinosa-Hernández, 2019) in that both emphasized religious ideology over other aspects of religiosity. However, the doctrinally religious class is distinct from a privately religious class because the doctrinally religious are ideologically—but not personally identifying as—religious or spiritual.

We did not find evidence in the main effects model that religious transitions were associated psychological well-being in later life. However, the significant interaction between staying strongly religious and gender revealed benefits of remaining strongly religious for men's psychological well-being in later life. The question of why lifetime religious stability is more beneficial for men's than for women's psychological well-being requires some informed speculation. A typical explanation relates to the role of social integration and community membership that religion offers and the deep relationship building that a life-time of religious involvement provides. Older men are less likely than women to have social ties outside their spouses. Additionally, the elevated status of men in many conservative religions may also confer advantages to well-being. It is possible that the ideological coherence offered by long-held religious beliefs and practices affords existential meaning to the post-retirement years of men. Such gender differences are also found societally and within ethnic subcultures. For example, Krause and Heywood (2013) found in a Mexican-American sample that church-based social support was more likely to bolster feelings of congregational belonging among older men than among older women. In national U.S. data, Mukerjee and Venugopal (2018) found that the positive association between religious attendance and self-rated health weakened more among women than it did among men. We speculate that social involvement that comes from maintaining religious participation provides psychological benefits for older men because older men's social networks are generally more attenuated than the social networks of older women.

These results should be interpreted in light of several limitations in the data. First, because of restrictions in sample size, we did not consider intermediate waves in the analysis. Although religiosity tends to be stable across the life course (Silverstein & Bengtson, 2018), some adults may change aspects of their religiosity between young adulthood and early old age. Because of the long period of time between young adulthood and young-old—approximately 45 years—patterns of change are likely more complex than depicted. We recommend that future studies with larger and more diverse samples address whether heterogeneous period effects (i.e., simultaneous growth of secularism and conservative evangelicalism in the last few decades) characterize transitions of religiosity classes from young adulthood to later life. Second, we used a fairly narrow definition religiosity in our analysis because the LSOG does not measure other important aspects of religiosity such as spirituality and private prayer in early waves. In addition, although psychological well-being has long been considered to be multidimensional (Ryff & Keyes, 1995), our results did not support the independent consideration of positive and negative dimensions of the Bradburn Scale of Psychological Well-Being. Third, our sample contained predominantly white and middle-income baby boomers. In addition, although our sample was not geographically restricted, the LSOG began in 1971 as a regional study in southern California. Consequently, these socio-demographic factors may limit the generalizability of our findings. Finally, we are reluctant to generalize our results to other generations since it is possible that the impact of religiosity on well-being may be weaker in more contemporary cohorts. Accelerated aging of the population will pose unique challenges, such as the burden imposed on the healthcare system (LeRouge et al., 2014; Knickman & Snell, 2002). Research on psychological well-being is revealing as it is a powerful predictor of older adults’ chronic physical disease and longevity (Steptoe et al., 2015). Notwithstanding the limitations of our research, this study provided new insights into religiosity and psychological well-being among baby boomers over the life course. We found a general decline of religiosity among baby boomers over their life course. This decline, if mirrored in more recent cohorts, implies that the psychological benefits of a stable religious life will be conferred on an increasingly smaller segment of the population.

		Wave-1	Wave- 9
Model 1	1	2491.01	2471.15	2475.14	2600.41	2580.56	2584.55
Model 2	2	2163.04	2119.36	2128.14	1986.07	1942.38	1951.16
Model 3	3	2171.13	2103.62	2117.19	1974.24	1906.73	1920.30
Model 4	4	2202.10	2110.76	2129.12	1998.68	1907.34	1925.70
Model 5	5	2232.21	2117.04	2140.19	2029.55	1914.38	1937.53
Model 6	6	2491.01	2471.15	2475.14	2061.64	1922.64	1950.58

Item response probabilities	Strongly religious	Doctrinally religious	Weakly religious
	Religious attendance	0.75	0.16	0.05
	Religious intensity	0.99	0.49	0.26
Wave-1	Religion had a greater influence	0.93	0.69	0.22
	Children get religious instruction	0.99	0.91	0.36
	God exists as described in the Bible	0.80	0.70	0.07
	Religious attendance	0.73	0.09	0.03
	Religious intensity	0.93	0.29	0.01
Wave-9	Religion had a greater influence	0.97	0.71	0.03
	Children get religious instruction	0.99	0.75	0.18
	God exists as described in the Bible	0.94	0.54	0.05
Latent class probabilities	Strongly religious	Doctrinally religious	Weakly religious
Wave-1	0.40	0.27	0.33
Wave-9	0.32	0.28	0.40

Wave-1\Wave-9	Strongly religious	Doctrinally religious	Weakly religious
Strongly religious	0.55	0.27	0.18
Doctrinally religious	0.12	0.37	0.51
Weakly religious	0.13	0.11	0.76

	Psychological well-being Wave-9
Intercept	4.86 (1.11)	4.35***	5.75 (1.19)	4.79***
Class memberships
Staying doctrinally religious^b	0.05 (0.47)	0.10	−0.49 (0.77)	−0.64
Staying weakly religious^c	0.02 (0.46)	0.05	−0.40 (0.53)	−0.75
Increasing religiosity^d	−0.44 (0.54)	−0.80	−0.87 (0.63)	−1.36
Decreasing religiosity^e	−0.21 (0.32)	−0.64	−1.18 (0.47)	−2.51*
Covariates
Psychological well-being (Wave-1)	0.14 (0.06)	2.35*	0.15 (0.06)	2.58**
Age (Wave-1)	0.02 (0.03)	0.57	0.01 (0.03)	0.40
Women (Wave-1)	−0.15 (0.19)	−0.80	−1.03 (0.44)	−2.33*
White (Wave-1)	0.16 (0.39)	0.41	0.20 (0.39)	0.51
Religious affiliation (Wave-1)	0.02 (0.32)	0.07	−0.05 (0.32)	−0.16
Education (Wave-9)	0.05 (0.09)	0.62	0.03 (0.09)	0.34
Annual income (Wave-9)	0.04 (0.02)	2.05*	0.04 (0.02)	2.04*
Married/cohabitate (Wave-9)	0.69 (0.29)	2.37*	0.60 (0.29)	2.04*
Retired (Wave-9)	0.23 (0.20)	1.15	0.19 (0.20)	0.96
Interaction terms
Staying doctrinally religious × Women	-	-	0.77 (0.93)	0.82
Staying weakly religious × Women	-	-	0.53 (0.71)	0.74
Increasing religiosity × Women	-	-	0.58 (0.97)	0.60
Decreasing religiosity × Women	-	-	1.56 (0.62)	2.24*

Footnotes

Declaration of Conflicting Interests

The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The authors disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: The project described was supported by Award Numbers 61457 by John Templeton Foundation and R21AG064512 by National Institute on Aging.

ORCID iD

Woosang Hwang

Author Biographies

Woosang Hwang is a postdoctoral fellow in the Aging Studies Institute at Syracuse University. He received a doctoral degree of human development and family science at Syracuse University. His research interests lie primarily in family dynamics and well-being over the life course.

Xiaoyan Zhang is a PhD candidate in the Department of Human Development and Family Science at Syracuse University. Her research interests include determinants and influences of health and well-being across the life course. Her work has appeared in Journal of Family Psychology and The Journals of Gerontology, Series B: Psychological Sciences and Social Sciences.

Maria T. Brown is an assistant research professor in Syracuse University's School of Social Work and Aging Studies Institute and a fellow of the Gerontological Society of America. She is a social gerontologist who uses the life course perspective to research health disparities and the later-life experiences of socioeconomically disadvantaged individuals, women, racial, ethnic, and sexual minorities, and people with dementia and their caregivers.

Sara A. Vasilenko is an assistant professor of Human Development and Family Science at Syracuse University. Her research uses innovative methodology to examine the longitudinal and multidimensional factors that influence health and well-being across the lifespan. She has a particular expertise in sexual behavior in adolescence and emerging adulthood, including how multiple aspects of religion influence sexual behaviors.

Merril Silverstein is inaugural holder of the Marjorie Cantor Chair in Aging Studies at Syracuse University, and serves as professor in the Department of Sociology and the Department of Human Development and Family Science. He received his doctorate in sociology from Columbia University, after which he served on the faculty of the Leonard Davis School of Gerontology at the University of Southern California. In over 200 research publications, he has focused on aging in the context of family life, with an emphasis on intergenerational relations over the life course and international-comparative perspectives. He currently serves as principal investigator of the Longitudinal Study of Generations which has collected data from the same families between 1971-2021, and is co-originator of the Longitudinal Study of Older Adults in Anhui Province, China, which began collecting data in 2001 and has continued to 2021. He is a Brookdale Foundation fellow, a Fulbright senior scholar, a fellow of the Gerontological Society of America, and between 2010-2014 served as editor-in-chief of the Journal of Gerontology: Social Sciences. In 2019 he was awarded the Matilda White Riley Distinguished Scholar Award from the Section on Aging and the Life Course of the American Sociological Association.

References

Allport

G. W.

Ross

J. M

. (1967). Personal religious orientation and prejudice. Journal of Personality and Social Psychology, 5(4), 432-443. https://doi.org/10.1037/h0021212.

Bakk

Vermunt

J. K.

(2015). Step-3 tutorial# 1: Step-3 models with covariates, distal outcomes, and multiple latent variables. Statistical Innovations Inc. Retrieved from https://www.statisticalinnovations.com/wp-content/uploads/LGtutorial.Step3_.1.pdf.

Bolck

Croon

Hagenaars

. (2004). Estimating latent structure models with categorical variables: One-step versus three-step estimators. Political Analysis, 12(1), 3-27. https://doi.org/10.1093/pan/mph001

Bradburn

N M

(1969). The structure of psychological well-being. Chicago: Aldine.

Bradshaw

Kent

B. V

. (2018). Prayer, attachment to God, and changes in psychological well-being in later life. Journal of Aging and Health, 30(5), 667-691. https://doi.org/10.1177/0898264316688116

Bravo

A. J.

Pearson

M. R.

Stevens

L. E

. (2016). Making religiosity person-centered: A latent profile analysis of religiosity and psychological health outcomes. Personality and Individual Differences, 88, 160 − 1169. doi.10.1016/j.paid.2015.08.049 https://doi.org/10.1016/j.paid.2015.08.049

Collins

L. M.

Lanza

S. T.

(2010). Latent class and latent transition analysis: With applications in the social, behavioral, and health sciences. John Wiley & Sons.

Cornwall

Albrecht

S. L.

Cunningham

P. H.

Pitcher

B. L

. (1986). The dimensions of religiosity: A conceptual model with an empirical test. Review of Religious Research, 27(3), 226-244. https://doi.org/10.2307/3511418

Diener

. (1984). Subjective well-being. Psychological Bulletin, 95(3):542-575. https://doi.org/10.1037/0033-2909.95.3.542

10.

Diener

Oishi

Tay

. (2018). Advances in subjective well-being research. Nature Human Behaviour, 2(4), 253-260. https://doi.org/10.1038/s41562-018-0307-6

11.

Ellison

C. G.

Levin

J. S

. (1998). The religion-health connection: Evidence, theory, and future directions. Health Education & Behavior, 25(6), 700-720. https://doi.org/10.1177/109019819802500603

12.

Fetzer Institute. (1999). Multidimensional measurement of religiousness/spirituality for use in health research: A report of the Fetzer Institute/National Institute on Aging working group. Retrieved from https://fetzer.org/sites/default/files/resources/attachment/%5Bcurrent-date%3Atiny%5D/Multidimensional_Measurement_of_Religousness_Spirituality.pdf.

13.

Fry

P. S

. (2000). Religious involvement, spirituality and personal meaning for life: Existential predictors of psychological wellbeing in community-residing and institutional care elders. Aging & Mental Health, 4(4), 375-387. https://doi.org/10.1080/713649965

14.

Hackney

C. H.

Sanders

G. S

. (2003). Religiosity and mental health: A meta-analysis of recent studies. Journal for the Scientific Study of Religion, 42(1), 43-55. https://doi.org/10.1111/1468-5906.t01-1-00160

15.

Hall

D. E.

Meador

K. G.

Koenig

H. G

. (2008). Measuring religiousness in health research: Review and critique. Journal of Religion and Health, 47(2), 134-163. https://doi.org/10.1007/s10943-008-9165-2

16.

Idler

E. L.

(2021). Religion and aging in the global context of secularization: Patterns, processes, consequences. In Ferraro

K. F.

Carr

(Eds.), Handbook of aging and the social sciences (9th ed., pp. 255-268). Academic Press.

17.

Klemmack

D. L.

Roff

L. L.

Parker

M. W.

Koenig

H. G.

Sawyer

Allman

R. M

. (2007). A cluster analysis typology of religiousness/spirituality among older adults. Research on Aging, 29(2), 163-183. https://doi.org/10.1177/0164027506296757

18.

Knickman

J R

Snell

E K

(2002). The 2030 problem: caring for aging baby boomers. Health Services Research, 37(4), 849–884. https://doi.org/10.1034/j.1600-0560.2002.56.x

19.

Koenig

H. G.

Kvale

J. N.

Ferrel

. (1988). Religion and well-being in later life. The Gerontologist, 28(1), 18-28. https://doi.org/10.1093/geront/28.1.18

20.

Koenig

H. G.

Larson

D. B

. (1998). Use of hospital services, religious attendance, and religious affiliation. Southern Medical Journal, 91(10), 925-932. https://doi.org/10.1097/00007611-199810000-00006

21.

Koenig

H. G.

McCullough

M. E.

Larson

D. B.

(2001). Handbook of religion and health. Oxford University Press.

22.

Krause

Hayward

R. D

. (2013). Church-based social support and a sense of belonging in a congregation among older Mexican–Americans. Review of Religious Research, 55(2), 251-273. https://doi.org/10.1007/s13644-012-0054-5

23.

Leondari

Gialamas

. (2009). Religiosity and psychological well-being. International Journal of Psychology, 44(4), 241-248. https://doi.org/10.1080/00207590701700529

24.

LeRouge

Van Slyke

Seale

Wright

(2014). Baby boomers’ adoption of consumer health technologies: survey on readiness and barriers. Journal of Medical Internet Research, 16(9), e3049. doi.10.2196/jmir.3049

25.

Levin

J. S.

Chatters

L. M

. (1998). Religion, health, and psychological well-being in older adults: Findings from three national surveys. Journal of Aging and Health, 10(4), 504-531. https://doi.org/10.1177/089826439801000406

26.

Lun

V. M. C.

Bond

M. H

. (2013). Examining the relation of religion and spirituality to subjective well-being across national cultures. Psychology of Religion and Spirituality, 5(4), 304-315. https://doi.org/10.1037/a0033641

27.

MacKinlay

. (2014). Baby-boomers ageing well? Challenges in the search for meaning in later life. Journal of Religion, Spirituality & Aging, 26(2-3), 109-121. https://doi.org/10.1080/15528030.2013.811711

28.

Mak

M. C. K.

Han

Y. M. Y.

You

Jin

Bond

M. H

. (2011). Building life satisfaction through attachment to mother and beliefs about the world: Social axioms as mediators in two cultural groups. Mental Health, Religion & Culture, 14(3), 223-239. https://doi.org/10.1080/13674670903456455

29.

Maselko

Kubzansky

L. D

. (2006). Gender differences in religious practices, spiritual experiences and health: Results from the US general social survey. Social Science & Medicine, 62(11), 2848-2860. https://doi.org/10.1016/j.socscimed.2005.11.008

30.

Miller

A. S.

Stark

. (2002). Gender and religiousness: Can socialization explanations be saved? American Journal of Sociology, 107(6), 1399-1423. https://doi.org/10.1086/342557

31.

Mukerjee

Venugopal

. (2018). Religiosity and health through the decades: Is there a gender difference? American Journal of Health Promotion, 32(4), 1028-1041. https://doi.org/10.1177/0890117116687886

32.

Nylund-Gibson

Choi

A. Y

. (2018). Ten frequently asked questions about latent class analysis. Translational Issues in Psychological Science, 4(4), 440-461. https://doi.org/10.1037/tps0000176

33.

Park

N. S.

Lee

B. S.

Sun

Klemmack

D. L.

Roff

L. L.

Koenig

H. G

. (2013). Typologies of religiousness/spirituality: Implications for health and well-being. Journal of Religion and Health, 52(3), 828-839. https://doi.org/10.1007/s10943-011-9520-6

34.

Roof

W. C.

(2001). Spiritual marketplace: Baby-boomers and the remaking of American religion. Princeton University Press.

35.

Roof

W. C.

McKinney

(1987). American Mainline religion: Its changing shape and future. Rutgers University Press.

36.

Ryff

C. D.

Keyes

C. L. M

. (1995). The structure of psychological well-being revisited. Journal of Personality and Social Psychology, 69(4), 719-727. https://doi.org/10.1037/0022-3514.69.4.719

37.

Scotto-Rosato

Baer

J. C

. (2012). Latent class analysis: A method for capturing heterogeneity. Social Work Research, 36(1), 61-69. https://doi.org/10.1093/swr/svs006

38.

Silverstein

Bengtson

V. L

. (2018). Return to religion? Predictors of religious change among baby-boomers in their transition to later life. Journal of Population Ageing, 11(1), 7-21. https://doi.org/10.1007/s12062-017-9216-0

39.

Silverstein

Bengtson

V L

. (2019). Longitudinal Study of Generations, California, 1971,1985, 1988, 1991, 1994, 1997, 2000, 2005 [Data File and Codebook]. Retrieved from https://www.icpsr.umich.edu/web/NACDA/studies/22100.

40.

Steptoe

Deaton

Stone

A A

(2015). Subjective wellbeing, health, and ageing. The Lancet, 385(9968), 640–648. https://doi.org/10.1016/S0140-6736(13)61489-0

41.

St. George

McNamara

. (1984). Religion, race and psychological well-being. Journal for the Scientific Study of Religion, 23(4), 351-363. https://doi.org/10.2307/1385724

42.

Strawbridge

W. J.

Shema

S. J.

Cohen

R. D.

Kaplan

G. A

. (2001). Religious attendance increases survival by improving and maintaining good health behaviors, mental health, and social relationships. Annals of Behavioral Medicine, 23(1), 68-74. https://doi.org/10.1207/S15324796ABM2301_10

43.

Suitor

J. J.

Gilligan

Pillemer

Fingerman

K. L.

Kim

Silverstein

Bengtson

V. L

. (2018). Applying within-family differences approaches to enhance understanding of the complexity of intergenerational relations. The Journals of Gerontology: Series B, 73(1), 40-53. https://doi.org/10.1093/geronb/gbx037

44.

Twenge

J. M.

Sherman

R. A.

Exline

J. J.

Grubbs

J. B

. (2016). Declines in American adults’ religious participation and beliefs, 1972-2014. Sage Open, 6(1), 1-13. https://doi.org/10.1177/2158244016638133

45.

Vasilenko

S A

Espinosa-Hernández

(2019). Multidimensional profiles of religiosity among adolescents: Associations with sexual behaviors and romantic relationships. Journal of Research on Adolescence, 29(2), 414–428. https://doi.org/10.1111/jora.12444

46.

Verbit

M F

(1970). The components and dimensions of religious behavior: Toward a reconceptualization of religiosity. In Hammond

P E

& Johnson

(Eds.), American Mosaic (pp. 24–39). New York: Random House number.

47.

Vermunt

J. K.

Magidson

(2016). Upgrade manual for latent GOLD 5.1. Statistical Innovations Inc. Retrieved from https://www.statisticalinnovations.com/wp-content/uploads/UpgradeManual5.1.pdf

48.

Voas

Chaves

. (2016). Is the United States a counterexample to the secularization thesis? American Journal of Sociology, 121(5), 1517 − 11556. https://doi.org/10.1086/684202

49.

Wuthnow

(1998). After heaven: Spirituality in America since the 1950s. University of California Press.

50.

Yoon

D P

Lee

E K O

(2004). Religiousness/spirituality and subjective well-being among rural elderly Whites, African Americans, and Native Americans. Journal of Human Behavior in the Social Environment, 10(1), 191–211. https://doi.org/10.1300/J137v10n01_05

51.

Zautra

Beier

Cappel

. (1977). The dimensions of life quality in community. American Journal of Community Psychology, 5(1), 85-97. https://doi.org/10.1007/BF00884787

	Baby boomer (n = 392)
		Wave-1		Wave-9
Variables	Range	n (%)	M (SD)	n (%)	M (SD)
Age			19.29 (2.66)		64.29 (2.66)
Gender
Men		160 (40.8)
Women		232 (59.2)
Race
White		361 (92.1)
Black		4 (1.0)
Hispanic		14 (3.6)
Others		12 (3.1)
Education	1–8				5.61 (1.39)
Marital status
Married				270 (68.9)
Living with a partner				19 (4.9)
Separated/Divorced/Widowed				87 (12.5)
Single/Never married				16 (4.1)
Annual household income	1–21				10.35 (6.01)
Religious affiliation
Mainline protestant		100 (25.5)		35 (8.9)
Evangelical protestant		69 (17.6)		64 (16.3)
Catholic		79 (20.2)		37 (9.4)
Jewish		50 (12.8)		43 (11.0)
Mormon		15 (3.8)		14 (3.6)
Others		13 (3.3)		12 (3.1)
None		62 (15.8)		182 (46.4)
Retirement
Yes				177 (45.2)
No				215 (54.8)
Religiosity indicators
Religious attendance: high		142 (36.2)		107 (27.3)
Religious attendance: low		246 (62.8)		284 (72.4)
Religious intensity: high		240 (61.2)		151 (38.5)
Religious intensity: low		146 (37.2)		238 (60.7)
Religion had a greater influence: high		249 (63.5)		204 (52.0)
Religion had a greater influence: low		140 (35.7)		185 (47.2)
Children get religious instruction: high		300 (76.5)		232 (59.2)
Children get religious instruction: low		88 (22.4)		155 (39.5)
God exists as described in the Bible: high		203 (51.8)		182 (46.4)
God exists as described in the Bible: low		171 (43.6)		203 (51.8)
Psychological well-being	0–10		5.87 (1.80)		7.46 (2.02)

		Wave-1			Wave- 9
Model	Classes (n)	BIC	AIC	SABIC	BIC	AIC	SABIC
Model 1	1	2491.01	2471.15	2475.14	2600.41	2580.56	2584.55
Model 2	2	2163.04	2119.36	2128.14	1986.07	1942.38	1951.16
Model 3	3	2171.13	2103.62	2117.19	1974.24	1906.73	1920.30
Model 4	4	2202.10	2110.76	2129.12	1998.68	1907.34	1925.70
Model 5	5	2232.21	2117.04	2140.19	2029.55	1914.38	1937.53
Model 6	6	2491.01	2471.15	2475.14	2061.64	1922.64	1950.58