Catalysts of Connectedness: A Case for Greater Complexity in Religious Identity Formation Research

Religious identity research has predominantly investigated effects of discrete factors, despite many factors exercising interconnected effects on religious connectedness, resulting in a limited understanding of the mechanism influencing religious identity development. This study examined the mechanism underlying the religious identity development in Jewish young adults, also showcasing the benefits of bringing together a range of known catalysts for examination in a single analytic model. Informed by Bronfenbrenner’s bioecological model, data from a sample of 1712 young adults from the Gen17: Australian Jewish Community Study (2018) was used to estimate bivariate and OLS regression models including moderated mediation to examine the relationships between Jewish schooling, critical Jewish experiences, parental religious connectedness and young adult religious connectedness. Jewish schooling significantly affected young adults’ religious connectedness; without mediating effects of other critical Jewish experiences, however, Jewish schooling effects were negligible. Upbringing by parents with high religious connectedness had an intensifying effect, while parents with low religious connectedness had a diminishing effect on the association between Jewish schooling and young adult religious connectedness. Those raised by parents with high religious connectedness had higher religious connectedness than those raised by parents with low-to-moderate religious connectedness, regardless of Jewish schooling. In addition, having a high proportion of Jewish peers in one’s friendship network was the most powerful of the critical Jewish experiences in mediating the effect of Jewish schooling on religious connectedness. Parents and Jewish friendship networks play important roles in the development of young adults’ religious connectedness, which is only apparent with research approaches that acknowledge the complexity of the formation of religious connectedness. The enduring nature of these influences even into young adulthood has implications for scholars of religion as well as religious communities, as there may be greater gain from investment in agency-building in families and coreligionist friendship networks rather than outsourcing to program development by communal institutions.

reveal the structure of the mechanism" (Wunsch et al. 2010: 8). This article will review the literature by conceptual model, arguing studies of religious connectedness have not yet comprehensively captured the socialization process's complexity. Moderated mediation analysis is then proposed as a superior approach to examining religious connectedness, as it more meaningfully acknowledges the reality, in Gestaltian parlance, that an interconnected examination of the whole produces a greater understanding than discrete studies measuring the sum of its parts. Two research hypotheses together with their practical and theoretical bases will then be outlined. A series of analyses will then be presented, showcasing the kind of insights OLS regression makes possible. Discussion of salient insights, limitations and methodological implications form the article's conclusion.

Issues in Research on Religious Connectedness Development
Longitudinal bivariate regression analyses conducted on surveys of predominantly Christian populations have mostly investigated religious connectedness for its amelioration of antisocial behavior (Mason and Windle 2001), contribution to spiritual coping, mental wellness (Reynolds et al. 2014), moral development (Hardy et al. 2011a), and identity formation (Hardy et al. 2011b). Among the latter category, bivariate analyses have identified correlations between self-reported religious beliefs and practices of parents and their adolescent children (Flor and Knapp 2001), and young adults' self-reported religious beliefs and perceived as well as actual parental religious beliefs (Milevsky et al. 2008).
A three-decade systematic review of religious connectedness research (Hardy et al. 2019) revealed religious connectedness rarely functioned as an outcome variable in mediation studies (analyses examining indirect effects of a predictor upon an outcome variable) and rarely functioned as a dependent variable in moderation studies (analyses examining interaction effects, that is, within which boundaries variables exert positive or negative interactions upon the relationship between a given predictor and outcome variable) and moderated mediation studies (see below for explanation). A few exceptions are that parents' and daughters' religious beliefs were found to be mediated by daughters' perceptions of parental beliefs (Okagaki and Bevis 1999). Family relationship quality and parenting style dimensions such as those promoting autonomy (Myers 1996) were found to moderate the association between family and individual religious connectedness. Gender was found to moderate the relationship between perceived paternal warmth and emerging adult religiosity, an effect only significant for males, suggesting a moderated mediation (Stearns and McKinney 2020). Finally, parent-adolescent attachment was found to moderate the relationship between parental and adolescent religious connectedness, the former positively correlated with adolescent sons' religious connectedness also suggesting a moderated mediation (Kim-Spoon et al. 2012). Most of the aforementioned studies did not control for religion or religious denomination-nor were religious education, co-religionist friendship networks, or religious community involvement controlled for-despite scholarly awareness that these factors predict religious connectedness (King and Roeser 2009). Most of the past quantitative research on religious connectedness focused on Christians to the neglect of Jews.

Issues in Research on Religious Connectedness Development of Jews
Religious connectedness is the focus of this study, but it is one of many forms of Jewish connectedness, others being ancestry (Lugo et al. 2013), culture (Brenner 1998), and ethnicity (Schoem 1989), religion being intertwined with these latter forms of connectedness in ways that make Judaism different from most forms of Christianity. There is plentiful qualitative research on these latter forms of Jewish connectedness (Fishman 2000;Prell 2011), but religious and cultural measures dominate quantitative research on Jewish populations (see, for example Boxer et al. 2021), ancestry and ethnicity limited mostly to markers of Jewish identification for 'Jews of no religion' (Alper and Cooperman 2021). Earlier research by this author explored several of these forms of Jewish connectedness (Bankier-Karp 2020). Given that this paper argues for increased conceptual and methodological complexity in understanding the mechanism responsible for religious connectedness, only research and measures relevant in predicting religious connectedness are included.
Bivariate studies have identified many factors which correlate with religious connectedness among Jews, including parental endogamy (Lugo et al. 2013), parental attitude toward religion (Fishman 2007), Jewish friendship networks (Kadushin and Tighe 2008), Jewish youth movement involvement and camp attendance (Cohen and Ganapol 1998), year-long yeshiva programs in Israel (Graham 2014;Jacobson 2004), Jewish volunteerism (Cohen 2001), and Jewish philanthropy (Graham and Boyd 2016). None of these studies controlled for or appeared cognizant of other influential factors, including age, gender, marital status, location and socioeconomic status (Hartman and Sheskin 2011); physical and mental wellbeing (Koenig 2013); and growing up in a family with Holocaust survivors (Samardzic et al. 2021). This is problematic as claims of causality cannot be based on bivariate analyses of crosssectional data that do not consider other factors. Nor is it possible to understand the mechanism underlying religious connectedness when these variables are measured in isolation, given the inevitability of their interconnected contributions.
The second issue with research on Jewish religious connectedness development relates to studies employing multivariate analysis, often used to identify independent relationships between one or more factors and religious connectedness, while controlling for influences of other factors. Jewish schooling was found to predict religious connectedness, even when controlling for religious denomination (a proxy for religious upbringing, Graham 2014). Jewish friendship networks were found to predict religious connectedness, when controlling for family denomination, family upbringing, and Jewish schooling (Cohen and Veinstein 2011). Jewish youth movement and camp participation were found to exert a modest but significant effect on religious connectedness when controlling for parental upbringing, select rituals in childhood, and variables such as sociodemographic status (Hartman and Sheskin 2011;Keysar and Kosmin 2004). None of these studies considered the mediating role played by Jewish friendship networks or Jewish communal involvement. These studies, moreover, controlled for-but did not examine effects of-parental religious connectedness, nor did they measure how the latter may have interacted with Jewish schooling and other possible indirect contributors to religious connectedness. This is problematic as these studies do not uncover the possible interplay between these factors, failing to uncover complexity in the mechanism underlying religious connectedness.
The final issue with research on Jewish religious connectedness development manifests in studies with excellent conceptual framing and methodological design, but which focus on the impact of a single factor. Service-learning studies reported sustained effects of Israel involvement on the religious connectedness of those with and without extensive Jewish educational backgrounds (Chertok et al. 2009). Longitudinal studies of the ten-day Taglit/Birthright Israel (BTI) trip were also found to have significant long-term effects on participants' religious connectedness, compared with non-participant applicants, a decade or more post-participation (Saxe et al. 2017;Wright et al. 2020). Nonetheless despite volunteerism and BTI being highly social experiences-it is intuitive that people are motivated to participate with Jewish friends and may consequently form Jewish social networks-there is a lack of attention paid to the predictive role of Jewish friends, together with other Jewish critical experiences such as youth movement and other communal involvement (see also Hartman and Sheskin 2011).
From the aforementioned research into Jewish populations, the following factors have been identified as significantly correlating with religious connectedness: Jewish schooling, Jewish peer networks, youth movement involvement, Jewish community involvement, Israel involvement and parental religiosity. The paper argues that the mechanism influencing religious identity development is poorly understood due the failure to include a range of known factors within a single analytic process, so their interconnected effects may be examined.

Moderated Mediation Analysis as a Better Alternative
Built on the foundational work on moderation and mediation (Aiken et al. 1991;Baron and Kenny 1986;Holmbeck 1997), moderated mediation is an analytic technique combining moderation and mediation (Edwards and Lambert 2007). The development of the PROCESS macro for SPSS-25 (IBM Corp., Armonk, NY, USA), together with lucid explanation of its usage (Hayes 2013), substantially simplified an otherwise complex procedure and increased the popularity of this form of analysis. Moderated mediation is superior to less complex regression analyses, as it identifies not only whether one or more independent variables are correlated with a given dependent variable, but also how one or more independent variables may mediate one another in exerting effects on a given dependent variable, and when-that is under what conditions-moderating effects do and do not occur. Such an approach is necessary to address the aforementioned issues in the research on religious connectedness development-namely that statistical relationships have been identified between Jewish schooling, coreligionist peers, youth movement and communal involvement, parental religiosity, and religious connectedness-but these studies did 1 3 not take all these factors into consideration simultaneously. Moderated mediation enables the aforementioned variables to be brought together in a single model, their interconnectedness acknowledged and measured.

H1
That the direct effect of Jewish schooling on religious connectedness will be significant-namely, young adults who received more years of Jewish schooling would have higher levels of religious connectedness than those who received minimal to no Jewish schooling-but the indirect effects of schooling will be greater-meaning that the relationship between Jewish schooling and young adult religious connectedness will be mediated by critical Jewish experiences-Jewish friendship networks, youth movement, Jewish community, and Israel involvement-and this latter analysis will identify a more powerful range of effects on religious connectedness.

H2
That the relationship between Jewish schooling and young adult religious connectedness will be moderated by parental religious connectedness-namely, being raised by parents with high levels of religious connectedness would have a positive interaction with Jewish schooling's effect on young adult religious connectedness-but Jewish schooling and parental religious connectedness will also interact in influencing the critical Jewish experiences and young adult religious connectedness-meaning that parents with high levels of religious connectedness will have a positive, intensifying interaction with Jewish schooling's direct effects on young adult religious connectedness, but the analysis of the indirect effects on young adult religious connectedness will identify a wider range of effects on religious connectedness.
The practical and theoretical bases for these hypotheses require brief elucidation. That religious schooling shapes the development of religious connectedness is evident, given that engagement with religious role models, sacred texts and rituals is critical to the fostering of beliefs, behaviors and a sense of belonging. Theoretically, the notion of schools as facilitators of religious connectedness is consistent with the literature on religious education (Otto 1958(Otto [1917). It is intuitive, moreover, that the relationship between Jewish schooling and young adult religious connectedness is mediated by Jewish critical experiences (H1), given that participation in Jewish friendship networks, youth movements, Jewish communal and Israel-related activities are often choices to engage with people who share similar values, beliefs and interests. These experiences would result in feelings of connectedness consistent with the literature on social identity theory (Turner and Oakes 1986), and cultural capital (Bourdieu 2011). That parents influence their children's experiences of Jewish schooling, as well as the other critical Jewish experiences (H2), is also unmistakable, given that parents not only fund and facilitate these experiences, but also color their children's perceptions by sharing their views of these experiences. Theoretically, this parental influence on the development of children's religious attitudes fits with the general literature on upbringing (Bao et al. 1999) as well as specifically religious upbringing (Ravitch 2002).

Data and Sample
The Monash University Human Research Ethics Committee granted approval for this research (Project number 7176). This study uses a sample of data from the Gen17 Australian Jewish Community survey (henceforth Gen17, https:// www. monash. edu/ arts/ acjc/ resea rch-and-proje cts/ past-proje cts/ jewish-attit udinal-surve ys/ gen17). The largest and most recent survey of Australian Jewry with a range of sociodemographic and Jewish connectedness variables, this non-probability convenience sample included 8,621 self-identifying Australian Jewish adults, recruited via community organization member lists, referral and an open web survey. (The methodology appears in Appendix III of the report, accessible via the above link). For this study, the sample was limited to young adults living in Melbourne or Sydney, who indicated they had completed (but not necessarily graduated from) high school (n = 1712). Young adults are defined as being ages 18 to 35, based on prior research of this population (Graham and Markus 2018) and sub-group analysis revealing the differences in this age range to be insignificant (Bankier-Karp 2020). The vast majority of Australian Jews, as well as young adults, live in Melbourne and Sydney and these populations have previously been deemed sufficiently similar for comparative analysis (Graham and Markus 2018). Young adults living in other states and territories, however, were omitted from this study, to eliminate the potentially confounding effects of living in smaller communities (Boxer 2013).

Measures
There is no consistent, agreed upon definition of Jewish connectedness (Cohen and Eisen 2000), so the author developed scales (see DeVellis 2017) to measure young adult and parental Jewish connectedness based on a thorough literature review and qualitative study of Jewish young adults in Australia (Bankier-Karp 2020). A factor analysis with oblique rotation (direct oblimin) resulted in five factors with loadings above 0.4 which, in combination, explained 65% of the variance (Table 1). The first of these factors represents young adult religious connectedness (Cronbach's alpha, α = 0.91)-a claim supported by the scholarship on Jewish religious connectedness-and was used as the dependent variable in this study.
A second factor analysis resulted in two factors with loadings above 0.4 which, in combination, explained 67% of the variance ( Table 2). The second of these factors represents parental religious connectedness (α = 0.74)-also theoretically supported-and was used as an independent variable in this study.

Independent Variables of Interest
The independent variables of interest measured Parental religious connectedness (α = 0.80), Jewish schooling (α = 0.80), Jewish friendship networks and involvement in Jewish youth movements, the Jewish community (α = 0.76), and Israel (α = 0.74). Jewish schooling was the primary independent variable of interest, Parental religious connectedness the moderator, Jewish friendship networks and the three forms of involvement proposed as mediators. Table 3 details descriptive statistics for the study variables.

Covariates
Nine sociodemographic variables identified earlier as influencing religious connectedness were controlled for in this study: age, sex, legal marital status, state of residence, country of birth, whether respondents had Holocaust survivor relatives, and three measures of wellbeing-financial, physical, and mental (Table 3).

Analytic Strategy
Four distinct lines of inquiry were followed to investigate the study hypotheses. The rationale underlying the components, structure and sequencing of the following analyses were derived from the relevant literature, and a qualitative study on Australian Jewish young adult identity (Bankier-Karp 2020).
1. The direct effect of Jewish schooling (X) on young adult religious connectedness (Y) was tested using a bivariate regression (Fig. 1). 2. The indirect effects of Jewish schooling (X) on young adult religious connectedness (Y) were tested using OLS regression The proposed mediator-critical Jewish experiences-consisted of four variables examined together in a serial multiple mediation model (SMMA). In the SMMA, young adult religious connectedness (Y) was operationalized as the dependent variable and Jewish schooling (X) as the independent variable. Jewish friendship networks (M 1 ), Youth movement involvement (M 2 ), Jewish community involvement (M 3 ), and Israel involvement (M 4 ) were operationalized as mediators (PROCESS macro model 6) (Fig. 2). The first and second statistical procedures operationalized H1. 3. The effects of parental religious connectedness (W) were also tested using OLS regression. The moderation analysis tested the interaction between Jewish schooling (X) and parental religious connectedness (W) on young adult religious connectedness (Y), predicting young adult religious connectedness from Jewish schooling, parental religious connectedness and their product (PROCESS macro model 1) (Fig. 3). 4. The effects of parental religious connectedness on Jewish schooling as well as the critical Jewish experiences on young adult religious connectedness were tested using moderated mediation analysis, in which the intervening variables  or mediators-Jewish friendship networks (M 1 ), Youth movement involvement (M 2 ), Jewish community involvement (M 3 ), and Israel involvement (M 4 )-were conceptualized as the mechanism through which Jewish schooling (X) influenced young adult religious connectedness (Y), a third variable called a moderator-Parental religious connectedness (W)-interacting with the pathways between the independent, mediator and dependent variables to contribute to the variation in Y. The mediation analysis was conceptualized as a partially mediated model c Jewish schooling scale: 'For how many years did you attend a Jewish Primary school?', and 'For how many years did you attend a Jewish Secondary school?' (Did not attend a Jewish school = 0 to seven years = 7) d Jewish friendship network: 'How many of your close friends are Jewish?' (None = 0, Less than half = 1, About half = 2, More than half = 3, All = 4) e Youth movement involvement: 'For how many years did you attend an Australian Jewish youth movement?' (I did not attend = 0 to three years = 3) f Jewish community involvement scale: 'How connected do you feel to Jewish communal life"' (Very unconnected = 0, Somewhat connected = 1, Somewhat connected = 2, Very connected = 3), 'How important is volunteering for charitable causes to your own sense of Jewish identity?', 'How important is belonging to a Jewish community to your own sense of Jewish identity?', 'How important or unimportant is donating money to charity to your own sense of Jewish identity?' (Extremely unimportant = 0, Fairly unimportant = 1, Fairly important = 2, Extremely important = 3) g Israel involvement scale: 'Do you consider yourself to be a Zionist?' (No = 0, Yes = 1), 'When international events put Israel in danger, which one of the following best describes how you feel?'(I do not feel any different about it than I would if another important foreign country were in the same sort of danger = 0, I feel rather more concerned than if another country was in the same situation = 1, I feel a special alarm because it is Israel and not some other country in danger = 2, My reaction is so strong that it is almost the same as if my own life was in danger = 3), 'Do you feel a responsibility to ensure that the State of Israel continues to exist?'(Strongly agree = 3, Tend to agree = 2, Tend to disagree = 1, Strongly disagree = 0), 'How important is supporting Israel financially to your own sense of Jewish identity?', 'How important is visiting Israel to your own sense of Jewish identity?' (Extremely unimportant = 0, Fairly unimportant = 1, Fairly important = 2, Extremely important = 3) h Holocaust survivor relatives: 'Were you, your mother, father, or grandparents survivors of the Holocaust? A survivor is defined as a person who experienced one or more of the following: During WW2, lived in, or escaped from, a country that was under Nazi rule or occupation, under the direct influence or control of the Nazis, or in front of the invading German army; In a concentration camp; In a ghetto (or similar place of incarceration in accordance with the German Slave Labor Program; In hiding or living under false identity/illegality for a period of at least 6 months in a Nazi-occupied or Axis country i Financial wellbeing: 'Which of the following terms best describes your current financial circumstances?' j Physical wellbeing: 'How is your health in general?' k Mental wellbeing: 'Please indicate which statement best describe your own state of health-in terms of anxiety and depression-today  with moderation of the direct effect and the first stage of an indirect effect across the four mediators (PROCESS macro model 8) (Fig. 4).
The third and fourth statistical procedures operationalized H2. For the second, third and fourth statistical procedure, the aforementioned sociodemographic factors were included as covariates. For the third statistical procedure, Post-hoc testing including simple slopes analysis (sample mean and plus and minus one standard deviation from the mean representing moderate, high and low parental religious connectedness, respectively, Hayes 2013) and the Johnson-Neyman technique (used to identify region(s) of significance along the continuum of the moderator where the conditional effect of X on Y changes from significant to not significant, Spiller et al. 2013), were used to understand the parameters of the detected effect.

Bivariate Regression Analysis
The bivariate regression was significant, with increases in Jewish schooling associated with increases in young adult religious connectedness (β = 0.24, F[1, 1711] = 106.90, p < 0.001). Jewish schooling explained 6% of the variance in young adult religious connectedness (Table 4).

Mediation Analysis
Mediation analysis revealed that as Jewish schooling increased, it had a significant direct effect on young adult religious connectedness (c' = 0.06, p = 0.020) (Fig. 5, the c path and Table 5). Consistent with hypothesis 1, Jewish schooling was positively associated with two of the four critical Jewish experiences-the more years of Jewish schooling received, the more likely it was young adults would report higher values across Jewish friendship networks and youth movement involvement, but not Jewish community and Israel involvement (Fig. 5, the four a paths). Partial serial multiple mediation analysis revealed, however, that all four critical Jewish experiences were positively related to higher young adult religious connectedness, the effects of all four experiences being statistically significant (Fig. 5, the four b paths and Table 5). Mediation analysis showed that Jewish schooling had a significant direct effect, as well as significant indirect effects-8 of the 15 indirect effect pathways being significant and the pathway between Jewish schooling to young adult religious connectedness via the Jewish peer network and Jewish community involvement having an effect more than double that of the direct effect (β = 0.14, p < 0.001, 95%CI 0.12-0.17)-and total effect (β = 0.32, p < 0.001, 95%CI 0.25-0.38) on young adult religious connectedness. These findings confirmed hypothesis 1 that the relationship between Jewish schooling and young adult religious connectedness was mediated by critical Jewish experiences-albeit only two of the four proposed variables. Moreover, not only were the indirect and total effects greater than the measured direct effect, but the examination of indirect effects identified a wider range of effects on religious connectedness.

Moderation Analysis
Moderation analysis revealed that the effect of Jewish schooling upon young adult religious connectedness was dependent upon parental religious connectedness (Fig. 6), the interaction accounting for 2.4% of the variance in young adult religious connectedness.
In terms of conditional effects of Jewish schooling on young adult religious connectedness, simple slopes analysis revealed parental religious connectedness significantly (p < 0.05) and negatively moderated the relationship between Jewish schooling and young adult religious connectedness at one point (SD − 1), the effect approaching zero as parental religious connectedness increased (conditional effects were 3.59, 9.73 and 16.00 at low, moderate and high values of parental religious connectedness, respectively, Fig. 7).     The Johnson-Neyman test revealed two ranges of significant values: 0 to 8.4 (negative effect) and 16.9 to 20 (positive effect) (Fig. 8). Post-hoc probing of the significant moderation analysis (Fig. 8) showed that Jewish schooling had a significant negative effect on young adult religious connectedness amongst those who reported low (β = − 0.20, p < 0.001) levels of parental religious connectedness during their upbringing, the effect ceasing to be significant (at the α level of significance) as parental religious connectedness increased.
Contrary to expectations, the association between Jewish schooling and young adult religious connectedness was not strengthened by a high level of the parental religious connectedness moderator. Further, this analysis revealed that being raised by parents with low religious connectedness had a diminishing impact on young adult religious connectedness, an effect which intensified even when the number of years of Jewish schooling increased. The association between Jewish schooling and young adult religious connectedness was dependent, therefore, on levels of the moderator. Despite these surprising interaction effects, however, the young adults raised by parents with high levels of religious connectedness had much higher levels of religious connectedness than those raised by parents with low religious connectedness, regardless of the number of years of Jewish schooling they received.

Moderated Mediation Analyses
Moderated mediation analysis revealed the conditional, direct and indirect effects of Jewish schooling on young adult religious connectedness ( Fig. 9 and Table 6). All moderating effects were significant, apart from Jewish community involvement. The conditional direct effect of Jewish schooling on young adult religious connectedness was significantly moderated by parental religious connectedness, the interaction accounting for 2.4% of the variance in young adult religious connectedness.
Consistent with Hypothesis 2, in terms of conditional effects of Jewish schooling on young adult religious connectedness, simple slopes analysis revealed parental religious connectedness significantly and positively moderated the relationship between Jewish schooling and young adult religious connectedness at one point (SD + 1). It also significantly and negatively moderated the relationship between Jewish schooling and young adult religious connectedness at another point (SD − 1), the effect approaching zero as parental religious connectedness approached moderate  Table 7).
The Johnson-Neyman test revealed two ranges of significant values: 0.00 to 10.53 (negative effect) and 17.81 to 20.00 (positive effect). Post-hoc probing showed Jewish schooling had a significant positive effect on the religious connectedness of young adults who reported high (β = 0.10, p < 0.001) levels of parental religious connectedness, and a significant negative effect on the religious connectedness of young adults who reported low (β = − 0.28, p < 0.001) levels of parental religious connectedness, these effects ceasing to be significant (at the α level of significance) as parental religious connectedness approached moderate levels. Being raised by parents with high religious connectedness had an intensifying effect on young adult religious connectedness and being raised by parents with low religious connectedness had a diminishing effect on young adult religious connectedness, effects which intensified with more years of Jewish schooling. The results echoed earlier research that parental religious connectedness was a significant moderator; the negative effect of being raised by parents with low levels of religious connectedness, however, was not in line with expectations. It is also worth noting that those who reported being raised by parents with high religious connectedness, had higher religious connectedness than those raised by parents with low to moderate religious connectedness across all levels of Jewish schooling (Fig. 7). The second hypothesis was confirmed by the finding that when examining the interaction of parental religious connectedness with Jewish schooling's direct and indirect effects on young adult religious connectedness, the inclusion of the critical Jewish experiences reveals not only the variability in parental effect, but also that parental religious connectedness has a wider range of indirect effects on religious connectedness than were apparent when merely examining direct effects.

Discussion
The main argument in this study is that the current understanding of religious identity development is limited by the methods utilized in its analysis. This argument is supported by the employment of four statistical procedures, whose results demonstrated how methods of increasing complexity may enhance the understanding of factors influencing religious connectedness. While bivariate regression-the first procedure-confirmed that young adults who receive more years of Jewish schooling report higher religious connectedness than those who receive minimal to no Jewish schooling, mediation analysis-the second procedure-confirmed the limited utility of merely measuring the direct impact of Jewish schooling on religious connectedness, revealing that greater effects occurred indirectly through other influencers-Jewish friendship networks and youth movement involvement (H1). Moderation analysis-the third approach-also shed light on the limited nature of studies which ignore not only parental impact but also the effects of being raised by parents with varying levels of religious connectedness. The moderation analysis in this study revealed that the association between schooling and religious connectedness was conditional upon parental religious connectedness; in other words, parents with low religious connectedness exercised a significant diminishing effect on Jewish schooling's impact on their offspring's religious connectedness. Compellingly, when young adults were raised by parents with moderate or high religious connectedness, such parental religious connectedness had no significant impact upon Jewish schooling's effect on their religious connectedness. When moderated mediationthe fourth statistical procedure used in this study-was conducted to examine additional effects of parental religious connectedness, the additional insights it yielded revealed the limited nature of the moderation analysis, the third statistical procedure  of this study. Moderated mediation analysis enabled all these insights to be examined in a single model, confirming not only that, for young adults, parents with low religious connectedness significantly decreased Jewish schooling's effect on their young adult religious connectedness, but that being raised by parents with high religious connectedness significantly increased Jewish schooling's effect on their religious connectedness (H2).

Parents are the Dimmer Switch of Their Children's Future Religious Connectedness
For young adults, being raised by parents with high religious connectedness exerted an intensifying effect on Jewish schooling's impact on their religious connectedness, findings supported by general studies reporting that parental consistency vis-à-vis religion is associated with higher religiosity in their offspring (Bader et al. 2006) and Jewish research reporting the positive effects of parental enthusiasm and rolemodelling on their children (Ravitch 2002). Being raised by parents with low religious connectedness exerted a diminishing effect on Jewish schooling's impact on their religious connectedness, a finding supported by general studies reporting that parental inconsistency vis-à-vis religion negatively affected children's religious connectedness (Bao et al. 1999) and Jewish qualitative research reporting negative effects of parental ambivalence on the religious connectedness development of children (Fishman 2007). What makes these finding so striking is that young adults raised by parents with high religious connectedness had higher religious connectedness than those raised by parents with low religious connectedness irrespective of the number of years of Jewish schooling they received (see Fig. 10). In addition, in the majority of cases, the diminishing effect of low parental religious connectedness on Jewish schooling's impact on young adult religious connectedness was present in the very same young adults whose parents who also paid for them to receive thirteen years of private Jewish day school education. It may be that for parents with low religious connectedness, providing a Jewish education for their children was motivated more by hazy notions of Jewish identity, or the Australian Jewish day schools' academic reputation than the Jewish studies classes offered. It is also possible that inconsistencies between home and school, if not overt antipathy towards the subject matter of Jewish studies classes, exerted an increasingly negative effect upon religious connectedness with more years of Jewish schooling.
What is also evident is that parents' impact does not end when they pay for their children's religious education. The findings of the moderation analysis challenge the claim that the greatest investment parents can make is sending their children to the schools and milieux which do the real work of fostering their religious connectedness (Keysar et al. 2000). The results also challenge studies which compare parents and peers to examine their respective influence (Harris 2011;Schneider 2010), instead suggesting that parents' levels of religious connectedness will predict whether their children will even have strong Jewish friendship networks in young adulthood. It therefore follows that the variability in young adult religious connectedness may be attributable in part to inherent differences in the young adults' experiences of being raised by parents with very different levels of religious connectedness. Despite claims that insight into predictors of religious connectedness might come from disentangling parental religious connectedness from other factors (Schoenfeld 1998), this study suggests that precisely by examining their interrelatedness, powerful insight is derived. The findings of the moderated mediation analysis are supported by Bronfenbrenner's (1979) claim that the understanding of phenomenon requires its multiple settings to be included in the research.

Jewish Friendship Networks are an Important Domino in Predicting Young Adult Religious Connectedness
This study also found that having Jewish friendship networks is the most powerful of the critical Jewish experiences in mediating the effect of Jewish schooling on religious connectedness. Of the fifteen indirect pathways by which Jewish schooling was hypothesized to influence religious connectedness in the mediation analysis, Jewish friendship networks had the largest effect. In addition, when measured directly, Jewish friendship networks have a stronger effect than even Jewish schooling on young adult religious connectedness. However, there was an even stronger indirect effect when the impact of Jewish schooling upon religious connectedness-mediated by Jewish friendship networks-was examined. An implication of this finding is that schools appear able to continue exerting influence on former students through friendships-and by fostering the value of having such friendshipswith co-religionists sharing common religious values. While the importance of the social is assumed in many recent studies (Chertok et al. 2007;Hartman and Sheskin 2011;Saxe et al. 2017;Wright et al. 2020), the use of moderated mediation analysis extends understanding of how Jewish friendship networks contribute to the effects of Jewish schooling on young adult religious connectedness. Given Bronfenbrenner's (1979) warning about the effects of the nested systems in which people find themselves, it is assumed that the measures and findings of this study are of relevance only in the national setting in which the study was conducted. In addition, the cross-sectional nature of the study, together with the fact that the dataset constitutes a convenience sample of Jewish young adults from Australia's two main Jewish cities, means that generalizations cannot be made beyond the current sample without further research. The novel nature of these study scales, moreover, means their generalizability beyond the studied population is unknown. Sociodemographic factors which have been theorized as explaining this phenomenon of sustained parental and friendship network influence on young adult religious connectedness include the Australian Jewish norms of young adults undertaking tertiary studies at local universities, young adults often living with their parents until at least the completion of their undergraduate degrees, and the high concentration of Jewish populations in Australia's two largest cities (Bankier-Karp 2020). In addition, most Australian Jewish young adults indicated their current close peers included friends from their childhood (Graham and Markus 2018). A near-majority of the young adults were also living in the state where they received their schooling (Bankier-Karp 2020). In combination, these factors highlight a striking stability in Australian Jewish social and communal networks. Australian Jewish friendship networks, comprised of childhood school friends and preserved by minimal geographic movement, may be contributing to the sustained effects of Jewish schooling. The prevalence of these three factors, contrasted with the North American and British norms of young adults studying at universities far from the community of their upbringing, at least in theory enables greater contact with family, Jewish friendship networks and familiar sites of Jewish communal engagement. Jewish schooling may have fewer lasting effects in countries where Jews are more mobile. Bronfenbrenner's (1979) ecological systems theory illuminates the possibility that Australian Jewry benefits from these and possibly other macro-structural forces, which might be contributing to stronger family bonds and more enduring effects of upbringing, as well as friendship networks extending the effects of schooling into young adulthood. This study argues that the mechanism responsible for religious connectedness requires greater conceptual and methodological complexity to better understand the religious identity development process. Moderated mediation analysis, for the insights it offers into that process, has been identified as a superior approach for the examination of religious connectedness, one which more meaningfully acknowledges the reality that an interconnected examination of the whole produces a greater understanding than discrete studies that measure the sum of its parts. This study, therefore, contributes to the literature by identifying mechanisms to measure already established relationships between Jewish schooling, religious connectedness and other factors. While Jewish connectedness is unique in its intermingling of religion, ancestry, culture and ethnicity, the majority of available data focuses on religious connectedness. As such, this paper does help correct a gap in the literature, presenting an approach that is relevant to the study of religion in general. Parents and friendship networks play important roles in the development of young adults' religious connectedness, which is only apparent with research approaches that acknowledge the complexity of the formation of religious connectedness. The enduring nature of these influences even into young adulthood have implications for religious research as well as religious communities, with greater potential gain from investment in building the agency of families and coreligionist friendship networks as milieux for religious identity development, rather than relegating responsibility for these experiences to communal institutions. In other words, in-sourcing, rather than outsourcing, may be the ultimate catalyst of religious connectedness.