Intergenerational Similarity of Religiosity Over the Family Life Course

Abstract

Objectives:

Research consistently shows that parents influence children’s religiosity. However, few studies acknowledge that there is within-group variation in the intergenerational transmission of religiosity. In this article, we examine whether and how congruence in religiosity between generations changes over the family life course and identifies unique parent–child trajectory classes.

Method:

We used eight waves of data from the Longitudinal Study of Generations, including 1,084 parent–child dyads beginning in 1971 when the children were adolescents and young adults, followed up to 2005. Growth mixture models (GMM) were tested.

Results:

GMM revealed four temporal patterns: stable similar, child weakens, child strengthens, and child returns. Results showed that children who were married were more likely to be members of the child-returns class than members of the stable-similar class.

Discussion:

Results are discussed in terms of the utility of the separation-individuation process and the life-course framework for understanding intergenerational differences and their stability over time.

Keywords

religion families life course dyadic differences intergenerational

Research has consistently shown that religiosity influences social factors from demographic and familial behavior to health outcomes (Ellison & Levin, 1998; Koenig, King, & Carson, 2012; Lehrer, 1999; Waite & Lehrer, 2003) and that a person’s religiosity is strongly associated with that of their parents (Glass, Bengtson, & Dunham, 1986; Min, Silverstein, & Lendon, 2012; Myers, 1996). Regardless of whether intergenerational transmission of values occurs from parents to children or vice versa (R. Hoge, Petrillo, & Smith, 1982; Kuczynski & Parkins, 2007; Roest, Dubas, & Gerris, 2010; Rohan & Zanna, 1996), evidence is clear that religiosity is a value that is maintained through generations (Bader & Desmond, 2006; Bengtson, Putney, & Harris, 2013; Copen & Silverstein, 2007; Glass et al., 1986; Miller, 2005; Vollebergh, Iedema, & Raaijmakers, 2001; Whitehouse, 2004). However, less is known about the temporal variation in parent–child similarity over the family life course. Most studies of intergenerational religiosity have examined correlations between generations or average trajectories in similarity over time. We suggest that there is likely heterogeneity in trajectories of religious similarity over time, with some parents and children demonstrating stability in their differences, others diverging due to the maturation and growing filial independence of adult children, and others converging as children acquire their parents’ roles such as becoming a spouse and/or parent.

This study addresses whether there are variations in trajectory patterns in religiosity between parents and children and whether these variations are attributable to life-course transitions such as marriage and parenthood. The literature discusses the concept of religiosity as being multifaceted, comprising religious identity, beliefs, and practices (Hackney & Sanders, 2003). Studies have examined various manifestations of religiosity ranging from belief systems, attendance at religious services, participation in religious organizations, religious problem-solving, and intrinsic religiosity (i.e., religion as an end; Bengtson et al., 2013; Dillon & Wink, 2007; Gullatte, Brawley, Kinney, Powe, & Mooney, 2010; Koenig, George, & Peterson, 1998; Koenig, Parkerson, & Meador, 1997). The current study considers religious intensity, beliefs, and attendance to represent as full a range of religiosity as possible in the data considered.

The Parent–Child Trajectories of Religiosity Over the Family Life Course

Previous studies discuss intergenerational transmission in the context of parents’ socialization of their children. Socialization theory suggests that parents have great influence on their children’s values and behaviors. Children acquire their parents’ values and attitudes through observational learning, modeling, parents’ direct coaching and training, and through classical and operant conditioning (Alwin, 1984; De Houwer, Thomas, & Baeyens, 2001; Smith & Self, 1980; Starrels, 1992). Based on socialization theory, intergenerational congruence in religiosity may manifest most strongly during early childhood and adolescence, when parents influence the religious orientations of their children by instituting formal religious training, celebrating religious holidays, attending religious services together, and stressing the importance of religious values in the home. Such congruence may become crystallized in that children maintain their received beliefs as they mature over time. The majority of studies have found significant parent–child religious congruence relatively early in the family life course, typically when the younger generation is in adolescence or emerging adulthood (Francis & Gibson, 1993; D. Hoge, 1994; Leonard, Cook, Boyatzis, Kimball, & Flanagan, 2013; Stolzenberg, Blair-Loy, & Waite, 1995; Wilson & Sherkat, 1994).

However, at older ages, it is possible that intergenerational congruence degrades over time, as the importance of other social agents (e.g., peers, teachers, and media) grows and children become more independent from their parents (Baltes, 1979; Erickson, 1992; Rossi, 1980; Settersten & Hagestad, 1996). As children move through the life course, family relationships and levels of investment in these relationships also change (Younnis & Smollar, 1985). As a result, young adults may begin to remove themselves from their parents’ sphere of influence and adopt their own identities and values (Thornton, Orbuch, & Axinn, 1995). This phenomenon has been discussed as a “second individuation process” that occurs between early adolescent to adulthood (Blos, 1967; Grotevant & Cooper, 1986; Smollar & Youniss, 1989). With regard to religion, Vermeer, Janssen, and Scheepers (2012) found that the church attendance of parents predicted the church attendance of adolescent children but not of adult children.

However, the separation and individuation process may occur while simultaneously maintaining connectedness with parents well into adulthood to aid in successful life transitions such as the completion of schooling and achieving financial independence (Younnis & Smollar, 1985). These findings argue for continued congruence in value orientations between generations.

Another perspective on value transmission proposes that children need to reach a specific life stage or a certain level of development before their parents’ influence manifests (Stolzenberg et al., 1995). As children transition into adulthood, they acquire roles that may align their values with those of their parents (Wilson & Sherkat, 1994). For example, as children acquire the role of spouse via marriage and as parent through childbirth, values or attitudes of parents and children may become more alike. From this perspective, it is possible that parent–child congruence may decrease as children mature into young adulthood but then increases as children enter into later adulthood and take on roles and responsibilities similar to those of their parents.

In summary, although evidence for religious congruence between parents and children is well established in the literature, the majority of research has been investigated from adolescence up until early adulthood (Dillon & Wink, 2007; Myers, 1996; Petts, 2009) and has not considered later periods when many life transitions occur. Furthermore, while some longitudinal studies have been conducted on this topic (Bengtson, Copen, Putney, & Silverstein, 2009; Min et al., 2012; Myers, 1996; Wilson & Sherkat, 1994), none to our knowledge have examined multiple waves of dyadic data over several decades of time nor have considered patterns change in parent–child religious similarity over time.

The current study adds to the literature by examining (1) clusters of trajectories in parent–child religious congruence and (2) the role of life-course transitions such as marriage and childbirth as factors that may narrow parent–child religiosity gaps over the life course. To address these questions, this investigation uses a unique longitudinal multigenerational data set collected over 30 years. We hypothesize that there will be existing distinctive patterns of parent–child trajectories. We also anticipate that life-course transitions such as getting married and becoming a parent will be associated with the probability of being in specific trajectory classes.

Method

Data and Sample

This study used parent–child dyadic data from seven waves of the Longitudinal Study of Generations (LSOG) to examine temporal patterns of parent and child religious trajectory over the adult life course. The LSOG began in 1971, with 2,044 respondents from 341 three-generation families. Grandparent (G1) participants were selected through multistage stratified random sampling of 840,000 individuals enrolled in Southern California’s first large Health Maintenance Organization. Enrollment into the survey was restricted to grandparents who had at least one grandchild between the ages of 16 and 26. All adult children (G2) and grandchildren (G3) in the designated age range descending from the G1s were recruited to participate in the survey. The sample was resurveyed in 1985, 1988, 1991, 1994, 1997, 2000, and 2005. Follow-up rates of eligible participants were 65% between 1971 and 1985 (with an additional 10% recapture rate in subsequent surveys) and averaged 85% for panels after 1985 (see Bengtson, Biblarz, & Roberts, 2002, for detailed information). All surveys were mail-back questionnaires, except in 2005, when about half the G3s participated via a web-based survey.

Given that this study aims to follow religiosity in children from adolescence to midlife to examine the extent of religious agreement between parents and children overtime, we relied on data¹ from G3 children (n = 554) and their G2 parents (n = 556) who participated in 1971. After matching children to their parents, the analytic sample consisted of 1,084 G2 parent–G3 child dyads consisting of 473 mother–daughter dyads (43.6%), 196 mother–son dyads (18.1%), 117 father–daughter dyads (10.8%), and 298 father–son dyads (27.5%).

Measures

In this study, religiosity was assessed with 3 items: religious service attendance, religious intensity, and conservative religious beliefs. Overall religiosity scores for each generation were calculated as the mean of those 3 items, ranging from 0 to 4. The reliability coefficient for the three dimensions was .72 for parents and .78 for children.
Religious service attendance was measured with a single item asking about frequency of attendance with response categories ranging from never (1) to everyday (6). We rescaled the range from 1 to 4 for comparability across dimensions.

Religious intensity was measured with the following question and response categories: “Regardless of whether you attend religious services, do you consider yourself to be not at all religious (1), somewhat religious (2), moderately religious (3), or very religious (4)?”

Conservative religious beliefs were measured by the strength of agreement with four statements representing conservative religious orthodoxy in terms of biblical literalism and religious universalism: (1) God exists as in the Bible, (2) We are all descendants of Adam and Eve, (3) The United States would be better if religion had more influence, and (4) Every child should receive religious instruction. Statements were evaluated (and scored) with the following response options: 1 = strongly disagree, 2 = disagree, 3 = agree, and 4 = strongly agree. A scale score was calculated by taking the mean of the 4 items. Reliability coefficients across generations and waves of measurement ranged from .84 to .89 (Glass et al., 1986).

Intergenerational congruence in each of the three dimensions of religiosity was operationalized as the raw difference between children’s religiosity scores and their parents’ religiosity scores in each wave of measurement between 1971 and 2005. The difference score ranging from −4 to 4, with zero indicating similarity, a positive value indicating that the parents’ religiosity is stronger, and a negative value indicating that the children’s religiosity is stronger.

Life-course transitions

We considered whether two family life transitions occurred by 1985: marriage (0 = not married; 1 = married) and the birth of first child (0 = no child; 1 = child birth). We chose 1985 as the threshold for family life transitions because this was the first wave of measurement after baseline when children averaged 33 years of age. At this wave, 80% of children had married and 73% had become parents.

Control variables included the gender of the children (0 = son; 1 = daughter) and the gender of the parents (0 = father; 1 = mother). Additional covariates include the age of the child, and the highest education achieved by both the children and parents, ranging from 1 to 8, with 1 = less than 8 years of education and 8 = postgraduate education. As an additional control, we considered religious affiliation (0 = non-Protestant; 1 = Protestant) to adjust findings for whether or not each generation was a member of the majority denomination in the sample.

Procedure

To identify discernible types of trajectories of intergenerational similarity in religiosity, we employed growth mixture modeling (GMM) and a variant of random effects modeling using Mplus 7 software (Jung & Wickrama, 2008). GMM groups heterogeneous patterns of individual change into common classes with each subject belonging to a specific growth pattern which then incorporates the ability to estimate individual predictors of class membership.

In order to identify the best functional form of the growth patterns, we successively examined a no-growth model (intercept only), a linear growth model (intercept and linear term), and a nonlinear growth model (intercept, linear term, quadratic term; a model with a cubic term was excluded, as the model failed to converge). The nonlinear model yielded the best fit across a variety of class specifications and was selected for further analysis.

To determine the number of classes that best describe the growth patterns in the data, model fit indices were compared across increasingly complex nonlinear models ranging from 1 to 5 classes. Typically, the most parsimonious model with an acceptable fit is considered optimal. There are a number of options for comparatively assessing fit, including the Akaike information criterion (AIC), the Bayesian information criterion, sample size–adjusted Bayesian (SSBIC) index, entropy, and Lo–Mendell–Rubin-adjusted likelihood ratio test (LMR-LRT). Models with lower values for the criterion indices and higher entropy values are considered better fitting models (Nylund, Asparouhov, & Muthen, 2007).

Models were estimated using a robust full information maximum likelihood (FIML) estimation procedure to treat missing data, based on the MAR assumption that attrition is random and conditional on observed variables in the model (Enders, 2001). Attrition of G3s was nonrandom with respect to gender and education, with greater attrition of men and lower educated individuals; both variables were included as predictors in the estimated models. In another diagnostic, we compared a model that included respondents with incomplete data to a model that considered only respondents who participated in all waves. There were no statistically significant differences in the estimates or in the general trends of trajectories.

Results

Table 1 shows characteristics of parents and children in this study. The average ages of parents and children in 1971 were 43.8 years and 19.2 years, respectively, and the average age of children in 2005 was 53.0 years. Slightly more than half of parents in the sample were mothers (54%) and a similar percentage (55%) of children were daughters. Approximately two fifths of parents (46.0%) and children (42.5%) in 1971 were self-described Protestants. A large majority (73%) of children became parents themselves by 1985.

Table 1.
Descriptive Characteristics of Parents and Children.

Variables M/% SD Range

Parent Mother (vs. father) 54.00 —

Age (Wave 1) 43.82 5.26 30–67

Education (Wave 1) 4.82 1.37 1–8

Protestant (Wave 1) 46.00 —

Religiosity (Wave 1) 3.03 0.86 1–4

Child Daughter (vs. son) 55.00 —

Age (Wave 1) 19.09 2.59 16–26

Education (Wave 2) 5.21 1.40 2–8

Protestant (Wave 1) 42.50 —

Religiosity (Wave 1) 2.69 0.91 1–4

Married (by Wave 2) 80.04 —

Became parent (by Wave 2) 72.73 —

Note. N = 1,084. Wave 1 = 1971; Wave 2 = 1985.

Goodness-of-fit statistics for the GMMs are provided in Table 2. These statistics suggest that the optimal number of trajectory classes was either four or five. Although the five-class model had slightly smaller AIC and SSBIC values than the four-class model which had smaller entropy as well, suggesting a worse model fit. In addition, one of the classes in the five-class model yielded a predicted membership probability of less than 2%, which represented only 18 parent–child dyads. Further, LMR-LRT also suggested that the four-class model was the better fitting model, the results from which conformed better with the theories reviewed in this article (Lynch & Taylor, 2016). Thus, the four-class model was selected as the more parsimonious and meaningful grouping of trajectories.

Table 2.
Fit Indices for GMM of Intergenerational Similarity in Religiosity.

Fit Indices One Classes Two Classes Three Classes Four Classes Five Classes

AIC 13,160.63 7,895.73 7,843.58 7,805.87 7,799.50

BIC 13,235.22 7,985.24 7,957.95 7,945.10 7,952.59

SSBIC 13,187.58 7,928.07 7,884.90 7,856.17 7,847.77

Entropy — 0.83 0.78 0.81 0.74

LMR-LRT — 180.72 (.01)* 60.42 (.33) 46.38 (.23) 68.06 (.07)

Note. AIC = Akaike’s information criterion; BIC = Bayesian information criterion; SSBIC = sample size–adjusted Bayesian information criterion; LMR = Lo–Mendell–Rubin-adjusted likelihood ratio test; GMM = growth mixture models.

Table 3 shows the intercept and linear and quadratic slopes for the four-class model. These estimates were used to illustrate predicted values in Figure 1, which displays growth patterns in similarity in religiosity between parents and children. The most common class, capturing 84.9% of the sample, was characterized by a stable-similar pattern that shows persistently high stability in intergenerational similarity (0 = completely similar) across time. The next largest cluster representing 7.3% of the sample was characterized as child weakens, characterized by high similarity at baseline but a decrease in similarity at an accelerating rate through young adulthood (B = .51, p < .001), before reaching a plateau of dissimilarity thereafter. This dissimilarity included higher parent religiosity than children. A third class capturing 3.3% of the sample exhibited a similar pattern to our second class, but the gap between parent and child resulted from higher religiosity in children than their parent: child strengthens. A fourth class capturing 4.4% of the sample exhibited a pattern of child returns across time. This class had similar religiosity as their parents at baseline but showed a decrease in similarity at an accelerating rate (B = −.63, p < .001), which eventually increased in similarity later (B = .10, p < .001).

Table 3.
Representation and Growth Estimates of Three Classes Derived From Growth Mixture Model.

Intercept Slope Quadratic

Class MP (%) Est. SE Est. SE Est. SE

Stable similar 84.9 .29*** .04 .01 .02 .00 .00

Child weakens (parent’s more religious) 7.3 .33* .17 .51* .07 −.05* .01

Child strengths (child’s more religious) 3.3 .07 .31 −.59* .12 .06* .02

Child returns 4.4 .01 .19 −.63* .10 .10* .01

Note. N = 1,084. Childs age at Wave 1 was controlled. MP = membership probability; Est. = standardized estimate; SE = standard error.

p < .05. ***p < .001.

Figure 1.
Trajectories of intergenerational similarity in religiosity over 34 years.

We found in supplemental analyses (not shown) both parents and children exhibited considerable stability in religiosity across time. Thus, it was not surprising that the stable-similar type of dyad was the predominant longitudinal pattern. Two divergent patterns (e.g., child weakens and child strengthens) correspond with the separation-individuation process, while the child-returns pattern is similar to the latency perspective that proposes that children need to reach a specific life stage before their parents’ influence on their values and attitudes is apparent.

Multinomial logistic regression examined factors that predicted membership in the child-weakens, child-strengthens, and child-returns growth classes relative to the stable-similar class that served as the reference group. Examining coefficients for the adoption of family roles related to marriage and parenthood, only marriage predicted class membership. The estimates shown in Table 4 reveal that married children were less likely to be in the child-strengthens class relative to the stable-similar class, suggesting that marriage induces stability in parent–child differences.

Table 4.
Multinomial Logistic Regression Estimates Predicting Membership in Intergenerational Religious Similarity Class.

Child Weakens (vs. Stable Similar) Child Strengthens (vs. Stable Similar) Child Returns (vs. Stable Similar)

Variables Est. SE OR Est. SE OR Est. SE OR

Parent

Mother (vs. father) 0.18 .21 1.20 0.65 0.37 1.92 −0.14 0.26 0.87

Age (Wave 1) −0.22* .16 0.80 −0.15 0.12 0.86 0.01 0.08 1.01

Education (Wave 1) 0.45* .19 1.56 0.18 0.22 1.19 0.15 0.11 1.16

Denom (ref. nonprotestant) −0.54 .68 0.58 −0.34 1.17 0.71 −0.68 1.01 0.51

Child

Daughter (vs. son) −1.43 .57 0.24 −0.92 0.80 0.40 0.40 0.67 1.49

Age (Wave 1) −0.35 .19 0.70 0.02 0.19 1.21 −0.04 0.10 0.96

Education (Wave 2) −0.37* .23 0.69 0.20 0.37 1.22 0.07 0.25 1.07

Married (by Wave 2) −0.52 .56 0.59 −1.10* 0.47 0.33 0.91 1.05 2.48

Became parent (by Wave 2) −1.34 .87 0.26 −0.30 0.79 0.74 1.27 1.40 3.56

Denom (ref. nonprotestant) −0.57 .82 0.57 −0.17 1.30 3.67 1.55 1.03 4.71

Note. N = 1,084. Wave 1 = 1971; Wave 2 = 1985; Est. = standardized estimate; SE = standard error; OR = odds ratio.

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

Discussion

The goal of this analysis was to examine the extent of religious agreement between parents and children over more than three decades in the family life course. Using longitudinal data between 1971 and 2005, we identified intergenerational differences in religiosity as children aged from adolescence and young adulthood into early middle age. Using a GMM approach, four distinct patterns of intergenerational similarity were identified: stable similar, child weakens, child strengthens, and child returns.

Consistent with previous studies, this investigation found that the large majority of parents and children remained on similar religious trajectories over time. This class could result when both parents’ and children’s religiosity remain stable (absolute stability) or the two groups increase or decrease in parallel (relative stability). To better understand these two scenarios, we further examined religious trajectories for parents and for children separately among the dyads in this class (not shown). We found that each generation remained in stable religious trajectories over time. This suggests that that this dominant class manifested highly stable absolute degrees of intergenerational religiosity even as children established independent identities and families of their own. This finding contradicts the dominant narrative of religious rebellion of younger generations and is more consistent with the family socialization theory, which suggests that early home life experiences influence not only the formation of beliefs, values, and attitudes but their maintenance as well.

One in 10 children experienced distancing from their parent’s religiosity shown in two divergent classes: child weakens and child strengthens. This group, while small, might be in the vanguard as the driving agent of social change observed in diverging religious orientations across generations. Our findings provide support for the power of later family events to tether families’ religiosity together. This tethering may be dependent on marital status, as being married was related to smaller intergenerational gaps. Two classes are similar in terms of existing gaps between parent and child over time but different in terms of the parents’ and children’s religiosity. We found that higher parental education is more likely in the child-strengthens class, which may suggest that religiosity in higher educated parents weakens more rapidly than their children resulting in relatively stronger religiosity among children than their parents later in life.

We found a relatively small class of dyads in which children returned to the religious orientations of their parents—a boomerang type characterized by early period divergence and later period convergence. This class exhibited an uptick in similarity as they entered midlife—a reversal toward greater similarity occurring roughly at the same age as when their parents expressed their religiosity at baseline. Evidence that becoming a parent is associated with greater similarity in religiosity with one’s own parents over time is consistent with the idea that new parents have strong desires to maintain the religious orientations of previous generations for the sake of their young children. Whether religiosity proceeds or is an outgrowth of normative family transitions remains an open question. Nevertheless, our findings speak to the potential interdependence of intergenerational religious linkages, marriage, and parental role adoption.

Our study focused on intergenerational religious differences and not the strength of religiosity per se. Based on descriptive statistics, however, we note that parents were significantly more religious than their children on all three measures, consistent with observations of progressively weakening religiosity across cohorts within American society. We also performed robustness checks to determine whether low religious dyads were responsible for our results. In dropping parent–child dyads in which both members of the dyad were not religious (4.7%), we found results consistent with analyses using the full sample. This result suggests that finding the large stable-similar class was not driven by secular parent–child dyads. It has been discussed in previous research that the intergenerational transmissions of nonreligiosity accompanied by the increase of the nonreligious youth is on the rise (Bengtson, Putney, & Harries, 2013). Future research will benefit from investigating different cohorts of parent–child dyads.

Finally, we examined an additive index of religiosity and did not present how trajectories were sensitive to diverse types of religious expression. Results for each of the three subdomains were generally similar, but percentages of class membership varied across them (results not shown). The model with conservative religious beliefs only showed a higher percentage of membership in the stable-similar class. Because beliefs represent a core aspect of religion, they may be more transmissible than the other religious dimensions. Therefore, it is perhaps not surprising that intergenerational concordance on this dimension would be most commonly observed. Future research will benefit from investigating between-domain heterogeneity in religious transmission.

Several limitations of the present research require mentioning. First, we examined only one cohort of intergenerational dyads; the large majority of whom were White non-Hispanic and originated from Southern California. The younger generation consisted of baby boomers who were first measured just following the societal upheavals and social changes of the 1960s, which might have exacerbated intergenerational differences and making our findings of strong religious consistency over time all the more surprising.

To the extent that race/ethnicity, geographic location, and cohort characteristics are uniquely associated with cross-generational family dynamics, caution should be exercised in generalizing the study’s findings. Second, although there is evidence of gender matching in religious transmission across generations, the interaction between the gender of the children and the gender of the parents could not be tested because including this interaction term resulted in model nonconvergence. Third, over 70% of our sample identified with a Judeo-Christian religion, mostly Protestant, Catholic, and Jewish denominations. However, questions measuring religious beliefs from Judeo-Christian scripture may not have been meaningful for the 30% of respondents without such affiliations. Thus, caution should be made in generalizing our findings beyond this particular religious tradition. Finally, although the extent of sample selection bias is mitigated by the FIML approach and the introduction of controls for baseline religiosity, it is important to note that the sample attrition rate from Wave 1 (1971) to Wave 2 (1985) was about 35%. It is therefore possible that a disproportionate number of more stable religious families remained in the sample.

Nonetheless, the limitations noted above should be weighed against the unique elements of the design features of the sample, including the availability of dyadic intergenerational data over a 34-year period that covers a good portion of the family life course. These features allowed for an assessment of children from adolescence to the early midlife of children to examine the roles played by early exposure to parents and later adoption of family roles during full adulthood.

In conclusion, our investigation revealed more stability than change in religiosity trajectories across generations. We postulate that this basic finding reflects the central role that religion plays in American families—an important linkage between younger and older family members that may supersede cohort and other developmental differences across generations. We hope this analysis has contributed to a better understanding of religiosity over the life course from an intergenerational perspective and the role that temporal and family dynamics play in the process.

Variables	M/%	SD	Range
Parent	Mother (vs. father)	54.00	—
	Age (Wave 1)	43.82	5.26	30–67
	Education (Wave 1)	4.82	1.37	1–8
	Protestant (Wave 1)	46.00	—
	Religiosity (Wave 1)	3.03	0.86	1–4
Child	Daughter (vs. son)	55.00	—
	Age (Wave 1)	19.09	2.59	16–26
	Education (Wave 2)	5.21	1.40	2–8
	Protestant (Wave 1)	42.50	—
	Religiosity (Wave 1)	2.69	0.91	1–4
	Married (by Wave 2)	80.04	—
	Became parent (by Wave 2)	72.73	—

Fit Indices	One Classes	Two Classes	Three Classes	Four Classes	Five Classes
AIC	13,160.63	7,895.73	7,843.58	7,805.87	7,799.50
BIC	13,235.22	7,985.24	7,957.95	7,945.10	7,952.59
SSBIC	13,187.58	7,928.07	7,884.90	7,856.17	7,847.77
Entropy	—	0.83	0.78	0.81	0.74
LMR-LRT	—	180.72 (.01)*	60.42 (.33)	46.38 (.23)	68.06 (.07)

		Intercept	Slope	Quadratic
Stable similar	84.9	.29***	.04	.01	.02	.00	.00
Child weakens (parent’s more religious)	7.3	.33*	.17	.51***	.07	−.05***	.01
Child strengths (child’s more religious)	3.3	.07	.31	−.59***	.12	.06***	.02
Child returns	4.4	.01	.19	−.63***	.10	.10***	.01

	Child Weakens (vs. Stable Similar)	Child Strengthens (vs. Stable Similar)	Child Returns (vs. Stable Similar)
Parent
Mother (vs. father)	0.18	.21	1.20	0.65	0.37	1.92	−0.14	0.26	0.87
Age (Wave 1)	−0.22*	.16	0.80	−0.15	0.12	0.86	0.01	0.08	1.01
Education (Wave 1)	0.45*	.19	1.56	0.18	0.22	1.19	0.15	0.11	1.16
Denom (ref. nonprotestant)	−0.54	.68	0.58	−0.34	1.17	0.71	−0.68	1.01	0.51
Child
Daughter (vs. son)	−1.43	.57	0.24	−0.92	0.80	0.40	0.40	0.67	1.49
Age (Wave 1)	−0.35	.19	0.70	0.02	0.19	1.21	−0.04	0.10	0.96
Education (Wave 2)	−0.37*	.23	0.69	0.20	0.37	1.22	0.07	0.25	1.07
Married (by Wave 2)	−0.52	.56	0.59	−1.10*	0.47	0.33	0.91	1.05	2.48
Became parent (by Wave 2)	−1.34	.87	0.26	−0.30	0.79	0.74	1.27	1.40	3.56
Denom (ref. nonprotestant)	−0.57	.82	0.57	−0.17	1.30	3.67	1.55	1.03	4.71

Footnotes

Authors’ Note

The study reported here is based on part of a dissertation by Joohong Min.

Declaration of Conflicting Interests

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

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This study was supported by the National Institute on Aging R56 AG007977.

Note

References

Alwin

D. F.

(1984). Trends in parental socialization values: Detroit, 1958–1983. American Journal of Sociology, 90, 359–382.

Baltes

(1979). Life span developmental psychology: Some converging observations on history and theory. In Baltes

Brim

(Eds.), Life span development and behavior (Vol. 2, pp. 225–279). New York, NY: Academic Press.

Bengtson

Biblarz

Roberts

(2002). How families still matter: A longitudinal study of youth in two generations. New York, NY: Cambridge University Press.

Bengtson

Copen

Putney

Silverstein

(2009). A longitudinal study of the intergenerational transmission of religion. International Sociology, 24, 325.

Bengtson

Putney

Harris

(2013). Families and faith: How religion is passed down across generations. New York, NY: Oxford University Press.

Bader

Desmond

(2006). Do as I say and as I do: The effects of consistent parental beliefs and behaviors upon religious transmission. Sociology of Religion, 67, 313–329.

Blos

(1967). The second individuation process of adolescence. The Psychoanalytic Study of the Child, 22, 162–186.

Copen

Silverstein

(2007). Transmission of religious beliefs across generations: Do grandparents matter? Journal of Comparative Family Studies, 38, 497–510.

De Houwer

Thomas

Baeyens

(2001). Association learning of likes and dislikes: A review of 25 years of research on human evaluative conditioning. Psychological Bulletin, 127, 853–869.

10.

Dillon

Wink

(2007). In the course of a life time: Tracing religious belief, practice, and change. Berkeley, CA: University of California Press.

11.

Ellison

Levin

(1998). The religion-health connection: Evidence, theory, and future directions. Health Education & Behavior, 25, 700–720.

12.

Enders

(2001). A primer on maximum likelihood algorithms available for use with missing data. Structural Equation Modeling, 8, 128–141.

13.

Erickson

(1992). Adolescent religious development and commitment: A structural equation model of the role of family, peer group, and educational influences, Journal for the Scientific Study of Religion, 31, 131–152.

14.

Francis

Gibson

(1993). Parental influence and adolescent religiosity: A study of church attendance and attitude toward Christianity among adolescents 11 to 12 and 15 to 16 years old. International Journal for the Psychology of Religion, 3, 241–253.

15.

Glass

Bengtson

V. L.

Dunham

C. C.

(1986). Attitude similarity in three generation families: Socialization, status inheritance, or reciprocal influence? American Sociological Review, 51, 685–698.

16.

Grotevant

H. D.

Cooper

C. R.

(1986). Individuation in family relationships. Human Development, 29, 82–100.

17.

Gullatte

M. M.

Brawley

Kinney

Powe

Mooney

(2010). Religiosity, spirituality, and cancer fatalism beliefs on delay in breast cancer diagnosis in African American women. Journal of Religion and Health, 49, 62–72.

18.

Hackney

Sanders

(2003). Religiosity and mental health: A meta–analysis of recent studies. Journal for the Scientific Study of Religion, 42, 43–55.

19.

Hoge

(1994). Vanishing boundaries: The religion of mainline Protestant baby boomers. Louisville, KY: Westminster John Knox Press.

20.

Hoge

Petrillo

Smith

(1982). Transmission of religious and social values from parents to teenage children. Journal of Marriage and Family, 44, 569–580.

21.

Jung

Wickrama

(2008). An introduction to latent class growth analysis and growth mixture modeling. Social and Personality Psychology Compass, 2, 302–317.

22.

Koenig

H. G.

George

L. K.

Peterson

B. L.

(1998). Religiosity and remission of depression in medically ill older patients. American Journal of Psychiatry, 155, 536–542.

23.

Koenig

H. G.

King

Carson

(2012). Handbook of religion and health. Oxford, England: Oxford University Press.

24.

Koenig

H. G.

Parkerson

G. R.

Jr Meador

K. G.

(1997). Religion index for psychiatric research. The American Journal of Psychiatry, 154, 885–886.

25.

Kuczynski

Parkin

(2007). Agency and bidirectionality in socialization: interactions, transactions, and relational dialectics. In Grusec

(Eds.), Handbook of socialization: Theory and research (pp. 259–283). New York, NY: Guilford Press.

26.

Lehrer

(1999). Religion as a determinant of educational attainment: An economic perspective. Social Science Research, 28, 358–379.

27.

Leonard

Cook

Boyatzis

Kimball

Flanagan

(2013). Parent-child dynamics and emerging adult religiosity: Attachment, parental beliefs, and faith support. Psychology of Religion and Spirituality, 5, 5–14.

28.

Lynch

Taylor

(2016). Trajectory models in aging research. In George

Kenneth

(Eds.), Handbook of aging and the social sciences (8th ed., pp. 23–51). New York, NY: Academic Press.

29.

Miller

(2005). Intergenerational transmission of religiousness and spirituality. In Miller

Dlaney

(Eds.), Judeo-Christian perspectives on psychology: Human nature, motivation, and change (pp. 227–244). Washington, DC: American Psychological Association.

30.

Min

Silverstein

Lendon

(2012). Intergenerational transmission of values over the family life course. Advances in Life Course Research, 17, 112–120.

31.

Myers

(1996). An interactive model of religiosity inheritance: Importance of family context. American Sociological Review, 61, 858–866.

32.

Nylund

Asparouhov

Muthen

(2007). Deciding on the number of classes in latent class analysis and growth mixture modeling: A Monte Carlo simulation study. Structural Equation Modeling, 14, 535–569.

33.

Petts

(2009). Trajectories of religious participation from adolescence to young adulthood. Journal for the Scientific Study of Religion, 48, 552–571.

34.

Roest

Dubas

Gerris

(2010). Value transmissions between parents and children: Gender and developmental phase as transmission belts. Journal of Adolescence, 33, 21–31.

35.

Rohan

Zanna

(1996). Value transmission in families. In Olson

Zanna

(Eds.), The psychology of values: The Ontario symposium (Vol. 8, pp. 253–276). Mahwah, NJ: Lawrence Erlbaum.

36.

Rossi

(1980). Aging and parenthood in the middle years. In Baltes

Brim

(Eds.), Life span development and behavior (pp. 138–207). New York, NY: Academic Press.

37.

Settersten

Hagestad

(1996). What’s the latest? Cultural age deadlines for family transitions. Gerontologist, 36, 178–188.

38.

Smith

Self

(1980). The congruence between mother’s and daughter’s sex role attitudes: A research note. Journal of Marriage and Family, 42, 105–109.

39.

Smollar

Youniss

(1989). Transformations in adolescents’ perceptions of parents. International Journal of Behavioral Development, 12, 71–84.

40.

Starrels

(1992). Attitude similarity between mothers and children regarding maternal employment. Journal of Marriage and the Family, 54, 91–103.

41.

Stolzenberg

Blair-Loy

Waite

(1995). Religious participation in early adulthood: Age and family life cycle effects on church membership. American Sociological Review, 60, 84–103.

42.

Thornton

Orbuch

Axinn

(1995). Parent-child relationships during the transition to adulthood. Journal of Family Issues, 16, 538–564.

43.

Vermeer

Janssen

Scheepers

(2012). Authoritative parenting and the transmission of religion in the Netherlands: A panel study. International Journal for the Psychology of Religion, 22, 42–59.

44.

Vollebergh

Iedema

Raaijmakers

(2001). Intergenerational transmission and the formation of cultural orientations in adolescence and young adulthood. Journal of Marriage and Family, 63, 1185–1198.

45.

Waite

Lehrer

(2003). The benefits from marriage and religion in the United States: A comparative analysis. Population & Development Review, 29, 255–275.

46.

Whitehouse

(2004). Modes of religiosity: A cognitive theory of religious transmission. New York, NY: Altamira Press.

47.

Wilson

Sherkat

(1994). Returning to the fold. Journal for the Scientific Study of Religion, 33, 148–161.

48.

Younnis

Smollar

(1985). Adolescents’ relations with mothers, fathers, and friends. Chicago, IL: University of Chicago Press.

		Intercept		Slope		Quadratic
Class	MP (%)	Est.	SE	Est.	SE	Est.	SE
Stable similar	84.9	.29***	.04	.01	.02	.00	.00
Child weakens (parent’s more religious)	7.3	.33*	.17	.51***	.07	−.05***	.01
Child strengths (child’s more religious)	3.3	.07	.31	−.59***	.12	.06***	.02
Child returns	4.4	.01	.19	−.63***	.10	.10***	.01