Promoting a Flourishing Society
 

Walking the Path: The Religious Lives of Young Adults in North America

Using Cardus Education Survey data, University of Notre Dame analysts say that attending a Protestant Evangelical school has a measurable effect on graduates that is distinct from the influence of family, socio-economic background, or church life.

January 30, 2018

School-sector research has shown the significant impact of attending a Protestant high school on religious outcomes in young adulthood. Graduating from an evangelical Protestant school has been associated with high levels of involvement in churches, including giving and volunteering. Compared to public-school graduates evangelical high schoolers on average are more supportive of traditional theological views, more committed to religious and spiritual obligations, more likely to see their life work in religious terms, and more likely to participate in devotional practices, such as Bible reading and prayer. The singular impact of the school effect is significant after other variables, such as family background are included.

One question that remains is the pathways through which evangelical Protestant (EP) high schools have a longer-term influence on religious beliefs and practices of their graduates. There are several explanations for the influence of EP schools on religiosity and religious orientations. Explanations that are defensible from previous research and possible to test using the Cardus Education Survey data include family and education variables. For example, EP schools are likely to focus their students on the family, so to speak, which may influence their family and marriage trajectories after high school. And these may in turn shape or reinforce the religious beliefs and practices instilled through EP school experiences and teachings. We would expect that EP schoolers are more likely to be married, and less likely to cohabit. They are likely to focus on raising children in the faith. By guiding graduates toward traditional marriage and family, evangelical Protestant schools may set up structural constraints and opportunities that encourage evangelical religious orientations and practices. Specifically, we would expect that EP-school graduates are more likely to attend church services regularly, and this tendency is enhanced and reinforced for them when they marry and have children. The mechanism is partly at the church level, in which family is closely linked to the structures and programs of evangelical churches. And it is generated in part as EP schoolers carry the culture of evangelicalism into their everyday life, which guides their family life. Evangelical culture and family structure mutually reinforce each other, and the result is higher levels of religious-service attendance. Yet it is also reasonable to assume that marriage and family in most Western societies would have an independent effect on religious outcomes—whatever the reason that brought the person to the altar or the baby shower. Religious ceremonies marking marriage and family transitions as well as societal expectations regarding what marriage and child-rearing is about may have a general influence toward heightened religiosity into young adulthood. Many a parent struggling to socialize unruly kids may find that religion is a good thing (at least in theory). For good reason (“God help us”), the frequency of prayer is a direct function of the number of children in the household. Less cynical is the view that marriage and children create powerful spiritual experiences, which could in turn shape the religious trajectories of adults. If EP schooling tends to channel students to marriage and family, the joys and challenges of each may lead to God—which would then explain at least part of the impact of EP schools on religiosity outcomes into adulthood.

A second major pathway that could explain the process through which EP schools influence their graduates over the longer term is the kind of college or university that EP students are more likely to experience. Most importantly, we know that EP schoolers are much more likely than other school-sector graduates to attend religious colleges and universities, especially the kind of evangelical college that is associated with the Council of Christian Colleges and Universities (CCCU) or with a similar organization in Canada, Christian Higher Education Canada (CHEC). Within the evangelical colleges, the orientations and teachings regarding religious beliefs and practices are likely to be developed and reinforced. The structure and culture of evangelical colleges and evangelical high schools point in similar directions. The result is that EP-school effects on religiosity and orientations may be significantly due to the fact that EP schools channel students into evangelical colleges and universities, which have their own impact on religion and religiosity into young adulthood.

In this report, we investigate the pathways between EP schooling and religious outcomes in young adulthood, focusing on the indirect effects of EP schooling that are mediated through family and education variables.

Data, Measures and Models 

The analysis for this report uses the U.S. and Canadian waves of the Cardus Education Survey (CES). This study was fielded in the U.S. with approximately 1,500 respondents in 2011, and the research design was replicated in 2014. The Canadian data was collected in 2012 and 2016. Each survey year included a large oversample of private-school attendees, resulting in about 2,000 private schoolers when the two waves are combined in each country. About 245 of these private schoolers were identified as having primarily attended an evangelical Protestant high school in the U.S., and 203 in Canada. Most of the survey questions were repeated in the later waves of the CES, and the variables were merged across years for this analysis.

Several variables regarding marriage and family were included in the survey. In this report, we use the current marital status of the respondent, focusing on whether the respondent is currently married and on whether the respondent is currently cohabitating. The survey also asked the respondents how many children they have, including biological and adopted children.

The survey included detailed questions on educational histories. In particular, the CES included a question on the name of the college or university attended by the respondent. The answers provided were used to place respondents into several college sectors, including public or government, private non-religious, Catholic, CCCU, and “other religious.” In Canada, we were able to create variables for CCCU or CHEC universities, and another for Catholic universities mentioned by the respondent.

To test the pathways through which EP-schooling influences various religious outcomes, we conducted regression analysis with multiple equations. This “path analysis” simultaneously estimated the impact of EP schooling on college sector, marital status, number of children, and each religious outcome. In each of the regressions, the comparison group for the EP-school sector is the public-school sector.

Using Mplus statistical software, we focused on total and indirect effects of EP schooling on the religious outcome of interest. That is, the results provided total effects of EP schooling, which included both the direct, or independent, effect of EP schooling on the religious outcome, as well as estimates of the extent to which EP schooling influenced family and education outcomes that in turn influenced the religious outcome. The total effect includes both the direct effect of EP schooling on the outcome (i.e., the effect that is not mediated through family or education variables), and the indirect effect that is mediated through the family and education variables included in the model. The indirect effect is limited by the extent that EP schooling influences marriage, family, and education; and the extent that these education and family variables in turn affect the religious outcome.

We include numerous control variables to isolate the EP school-direct and indirect effects on religion outcomes. In predicting the religious outcome, we control for gender, race and age. Also controlled are family-background variables, including whether parents were divorced or separated, and whether the respondent lived with their adopted or biological parents throughout their childhood. We controlled for religious background with variables for whether the respondent reported that their mother was conservative Protestant, and another variable for whether respondents reported that their mother was a traditional or orthodox Catholic. We included identical measures for fathers as well. The highest educational level of a parent was included as a control. The college/university sectors of non-religious private, Catholic, CCCU, and “other religious” were included as binary variables in the U.S. analysis. In the Canadian analysis, we included binary variables for evangelical university attendance and Catholic university attendance. A binary for those who did not complete a BA was also included. The college comparison group is the public university sector. When predicting the effect of EP schooling on number of children, we included all the demographic and family-background variables. This set of controls was also used in the regression predicting the variables for being married, being in a cohabitating relationship, and each of the college-sector variables, including the CCCU/CHEC binary variable. All of the effects discussed below are net or independent of the effects of the control variables.

Results

Religious Practices

U.S. Findings 

We begin with various religious-practice outcomes, since the indirect effects in these cases are more intuitive. First, the frequency of religious-service attendance. We can think of several reasons why religious-service attendance would not only be directly affected by EP schooling experiences and teachings but also why this effect would be mediated by the effect of EP schooling on marriage, family, and postsecondary educational experiences. EP schoolers are more likely to be married and to have children. The connection between marriage and evangelical churches, and the emphasis on integrating religion into marriage, would likely promote religious-service attendance. More children likely also places greater urgency on religious-service attendance as parents are concerned about aspects of child socialization and opportunities available through congregations. And we know that EP schooling emphasizes not only religious-service attendance, but also marriage and family. Thus we expect indirect pathways that run from EP schooling through marriage and family and ultimately influence religious-service attendance.

What we find is that EP schooling, not surprisingly, has a strong total effect and a strong indirect effect on religious-service attendance in young adulthood. The frequency of attendance at religious services is much higher on average among EP schoolers than among public schoolers, net of the controls for family background. Almost a third of the total effect of EP schooling is indirect—that is, it is due to the effect of EP schooling on marital status, number of children, and college sector. Since EP schooling influences whether respondents are married or not, for example, and marital status in turn influences religious-service attendance, we conclude that part of the reason that EP schooling affects religious-service attendance is because of its effect on the likelihood of being married. Altogether, we find that marriage and number of children are significant mediators of the relationship between EP schooling and religious-service attendance. This makes sense since it is reasonable that EP schooling would push graduates toward marriage, that marriage has an “elective affinity” with religious-service attendance, and that EP graduates would expect to take seriously religious obligations within marriage. In addition, we find that attending an “other religious” or CCCU college or university mediates the relationship between EP schooling and attendance. CCCU attendance accounts for the largest portion of the indirect effect—about 12 percent of the total effect of EP schooling on religious-service attendance. This seems to indicate that by channeling their graduates toward CCCU colleges and universities, EP schools indirectly enhance the religious-service attendance of their graduates.

Personal prayer is affected directly and indirectly by EP-school experiences. The total effect of EP schooling on the extent that respondents pray alone is very strong. Breaking that down, we find that the indirect effect of EP schooling on prayer—through family and education variables—adds up to about 21 percent of the total EP-schooling effect. The significant mediating variables in this case include the number of children and attending a CCCU college or university. About 9 percent of the mediated effect of EP schooling on prayer is due to CCCU attendance. Thus we conclude that EP schooling influences the practice of personal prayer in numerous ways, some of which we do not test directly in these models. What we are able to test—effects mediated by marriage, cohabitation, number of children, and college sector—reveals that number of children generates conditions that favour the practice of personal prayer. And attending a CCCU institution does the same. EP schooling influences both the number of children and the likelihood of attending a CCCU institution, and thus indirectly influences personal prayer through these mechanisms.

Reading the Bible is much more frequent on average among EP schoolers, even net of family-background controls. About 20 percent of this very strong EP schooling total effect is mediated through the family and education variables in the model. The key mediators that link EP-schooling and Bible reading are number of children and at tending a CCCU college or university. About 10 percent of the total effect of EP schooling on Bible reading is mediated by CCCU attendance. Again, EP schooling matters in many ways for boosting the practice of Bible reading, and, among these, we find evidence that family and an evangelical college explain part of the EP schooling effect on Bible reading.

Similarly, EP schooling is strongly related to reading other religious literature besides the Bible. About 31 percent of the total effect of EP schooling is indirect, mediated through three of our family and education variables. The significant mediators include number of children, attending an “other religious” college, and attending a CCCU institution. About 12 percent of the indirect effect is mediated by CCCU attendance. Again, one of the main reasons that we find for the influence of EP schooling on the reading of religious literature is the effect of EP-schooling on the likelihood of attending a CCCU institution.

Canada Findings 

The findings in Canada are very similar. When predicting frequency of religious-service attendance, for example, EP schooling has an extremely strong positive effect. This total effect is mediated by marriage, number of children, and attending a CCCU or, the Canadian equivalent, CHEC university. About 16 percent of the total effect of EP schooling is mediated by the education and family variables, which is a smaller indirect effect than we found in the U.S.

We find a strong positive effect of EP schooling on frequency of prayer in Canada. About 20 percent of the total effect of EP schooling is indirect. The indirect effect is mediated through marriage, number of children, and attending a CHEC or CCCU institution. These effects are very similar to the U.S. findings, except that marriage is an important mediator in Canada.

As in the U.S., we find a very strong effect of EP schooling on frequency of reading the Bible. About 10 percent of this effect is mediated, primarily through number of children and evangelical university attendance. In addition, the indirect effect through marriage is marginally significant. Although the mediators are the same, the indirect effect is much less than in the U.S.

EP schooling in Canada has a strong effect on reading other religious literature, and about 21 percent of this effect is indirect, mediated through number of children and evangelical university attendance. These findings are very similar to the U.S findings.

Religious Obligations

We next consider orientations to various religious practices. The CES included several questions on the extent to which the respondents believed that certain practices should be considered moral or religious obligations.

U.S. Findings 

The sense of obligation to participate regularly in spiritual practices, such as prayer and Bible reading, is strongly and positively related to EP schooling. The family and education variables mediate about 30 percent of the total effect of EP schooling on this view of spiritual-practices obligation. Number of children is the most important family variable that mediates this relationship. The strongest mediator, however, is attending a CCCU college or university, which accounts for nearly 14 percent of the total effect of EP schooling.

The EP schooling (total) effect is particularly strong in predicting the sense of a religious obligation to tithe 10 percent of income. About 24 percent of this effect is indirect. Again, number of children is the significant mediator among the family variables. The lower likelihood of cohabitating for EP schoolers is another important pathway, but this is not statistically significant. More importantly, attending an “other religious” or CCCU college or university is a strong mediator of the EP schooling effect. About 11 percent of the total effect of EP schooling on this outcome is mediated by CCCU attendance.

EP schoolers are much more likely to agree that they have an obligation to defer to the spiritual authority of church leaders. About 28 percent of this relationship is indirect, mediated through family and education variables. Number of children has a substantial and significant mediating effect in the relationship between EP schooling and obligation to church authority. By far the strongest mediating effect involves CCCU attendance. About 12 percent of the total effect of EP schooling on this outcome is mediated by CCCU attendance.

Canada Findings

In Canada, EP schooling has a very strong positive effect on a sense of religious obligation to take up spiritual practices. About 15 percent of this effect is indirect, mediated through marriage, number of children, and evangelical university attendance. Again, the mediators are the same on each side of the border, though the size of the indirect effect is less in Canada.

EP schooling has a strong positive effect on the sense of obligation to tithe. About 9 percent of this total effect is mediated by number of children and marriage.

There is a strong direct effect of EP schooling on the sense of obligation to defer to the authority of church leaders. But in Canada the indirect effect is small and statistically insignificant.

Integration of Faith and Everyday Life

U.S. Findings 

EP schoolers are much more likely to say that they hold a job that fulfills God’s calling in their lives. About 22 percent of this effect is mediated by the family and education variables. About 9 percent of the total effect is mediated by the higher number of children in EP-schooler families. The most important mediating variable, however, is attending a CCCU college or university. About 11 percent of the total effect of EP schooling on a sense of calling is due to the effect of EP schooling on the likelihood of attending a CCCU college or university. Again, we see Christian college and high school working in sequence to influence orientations to the integration of faith and life.

Similarly, EP schoolers are much more likely to say that God has called them to their particular line of work. About 25 percent of this total effect is mediated by family and education variables. Nine percent can be chalked up to number of children, and about 12 percent is mediated by attending a CCCU college or university.

The CES survey asked respondents how they tend to deal with difficult moral decisions in their lives. Of the several alternative responses, one emphasized the importance of God or Scripture in making tough moral choices. Moral decision-making in terms of God or Scripture is strongly related to EP schooling, and about 23 percent of this effect is mediated through the education and family variables. About 5 percent of this mediated effect is due to number of children, and about the same amount is due to being married. The strongest mediated effect is through CCCU attendance, which accounts for about 10 percent of the total effect of EP schooling on moral decision-making.

Canada Findings 

The Canadian sample reveals a very strong effect of EP schooling on the value placed on a job that fulfills God’s calling for the respondent’s life. About 17 percent of this total effect of EP schooling is indirect. In this case, the important mediating variables are marriage and number of children.

EP schooling has an extremely strong effect on the sense that God has called the respondent to their particular line of work. The key mediators here are number of children and evangelical university attendance. Marriage is marginally significant as a mediator in this model. About 13 percent of the total effect of EP schooling on a sense of calling is mediated through these variables.

That God or Scripture provide the most important guidance when making a difficult decision in life is extremely strongly affected by attending an EP school. About 15 percent of this total EP-school ing effect is mediated through marriage, number of kids, and attending an evangelical university.

Spirituality

We next consider several outcomes that are likely to be related to spirituality or personal religious orientations.

U.S. Findings

EP schooling has a strong positive total effect on the extent that individuals have a sense of gratitude. A significant though relatively small portion of this total effect is mediated through family and education variables. In particular, about 5 percent of this effect is due to the greater likelihood that EP schoolers will be married. Another 5 percent is due to the higher number of children in EP-schooler families. The EP-schooling effect is partially mediated through marriage and family in this case, but the family and education variables are clearly not the only pathway to higher levels of gratitude among EP schoolers.

EP schooling has a very strong and significant effect on the extent that the respondent believes that they can find spiritual peace even in the midst of life problems. About 38 percent of this total effect of EP schooling on spiritual peace is mediated by the education and family variables in the model. Marriage, number of children, “other religious” college, and CCCU college are significant mediators for the relationship between EP schooling and spiritual peace. About 14 percent of the total effect of EP schooling is mediated by attending a CCCU college or university.

On average EP schoolers are higher on a measure of the extent that they try to strengthen their relationship with God. About 30 percent of the total effect of EP schooling on this outcome is indirect. In particular, the number of children is a strong mediating variable for this relationship between EP schooling and strengthening one’s relationship with God. The other strong pathway is through CCCU college or university attendance, which accounts for about 10 percent of the total estimated relationship between EP schooling and trying to strengthen one’s relationship with God.

The extent that respondents report that they experience a deep communion with God is very strongly influenced by EP-school experience. Investigation of the pathways for this influence reveal that about 21 percent of this effect is mediated by our education and family variables. Each of the marriage and family mediators seems to provide a pathway to communion with God, but the only statistically significant pathway is through number of children. CCCU attendance does seem to matter as well, but this effect is not statistically significant.

CES respondents answered a question about the extent that they believe that everything that happens to them is part of God’s plan, including suffering. EP schoolers are much higher on average in agreeing with this statement. About 21 percent of the total effect of EP schooling on this “spiritual-plan” variable is due to indirect effects in our model. The most important pathways are through number of children and CCCU attendance.

The EP-schooler total effect is very strong in predicting the extent that respondents experience doubt about their religious faith and beliefs. EP schoolers report a lower level of religious doubt, and about 10 percent of that total effect is mediated through education and family variables. The number of children is the main pathway in mediating this relationship. Interestingly, attendance at religious colleges or universities does not significantly mediate the relationship between EP schooling and religious doubts.

EP schooling has a strong positive effect on the extent that respondents report that their spiritual experience is fulfilling. About 30 percent of the total effect of EP schooling on spiritual fulfillment is mediated by our education and family variables. Marriage, number of children, and CCCU attendance matter significantly in this case. About 10 percent of the total effect of EP schooling is due to CCCU attendance.

Canada Findings

In the Canadian sample, EP schooling does not have a significant total effect on a sense of gratitude. But there is a positive indirect effect that is mediated through marriage, number of children, and evangelical university attendance. This is one of the few instances in which the effect of EP schooling is entirely accounted for by our mediating variables.

EP schooling in Canada does have a strong effect on the extent that respondents experience spiritual peace. About 16 percent of this effect is indirect. The primary mediators are again marriage, number of children, and evangelical university experience.

EP schooling has a very strong effect on the extent that respondents report that they try to strengthen their relationship with God, and about 16 percent of this total effect is indirect. Again, the EP-schooling effect is mediated through marriage, number of children, and evangelical university attendance.

Similarly, a sense of communion with God is strongly affected by EP schooling attendance. About 15 percent of this effect is indirect. The key mediating variables are marriage and children, and, to a lesser extent, evangelical university.

The effect of EP schooling on the belief that everything that happens, including suffering, is God’s will is extremely strong. The indirect effect is about 12 percent of the total effect, and is primarily mediated through marriage and children.

In contrast, the negative relationship of EP schooling on the extent of religious doubt is not large and is not statistically significant. Interestingly, a small indirect effect of EP schooling on religious doubt is actually positive and is mediated through the variable for number of children.

The total effect of EP schooling on a sense of being spiritually fulfilled is large and significant. About 12 percent of this effect is indirect, mediated through children, evangelical university attendance, and, to a lesser extent, marriage.

Religious Beliefs

U.S. Findings

The CES survey asked respondents about their view of God, and offered several alternatives, including that God is a personal being involved in human lives today. That respondents believe that God is a personal being is strongly dependent on having attended an EP school. The mediated effect is less substantial here but still significant. About 21 percent of the total effect is mediated by our education and family variables. The most important mediators are number of children and attending a CCCU college or university. About 9 percent of the total effect is mediated by attending a CCCU institution.

The belief that Jesus Christ is the only way to salvation is incredibly strongly predicted by EP-schooling attendance, even net of the family-background variables. About 19 percent of this effect is mediated by family and education. In particular, number of children is a very strong and significant mediator of the EP-schooling and salvation-views relationship. The other strong mediating factor is CCCU attendance. About 9 percent of the total effect of EP schooling on salvation views is mediated by CCCU experience.

EP schoolers are much more supportive of the view that the Bible is infallible in matters of faith and practice. About 18 percent of the EP-schooler total effect is mediated through the education and family variables. Again, the number of children is a significant mediator in this relationship. Most impressive is the extremely strong mediating effect of CCCU attendance. Slightly more than 9 percent of the total effect of EP schooling on orthodox view of the Bible is mediated by CCCU attendance. Similar findings emerge regarding the view that the Bible has errors in science and history. EP schooling has a negative total effect on this outcome, and this effect is mediated by number of children and by attending a CCCU institution.

Canada Findings

In the Canadian sample, EP schooling has a moderate positive direct effect on the view that God is a person, but the indirect effect is small and not statistically significant. It appears, then, that EP schooling does in fact teach or reinforce an orthodox view of God, but that we cannot explain the EP-schooling effect through our marriage or education variables.

The view that Jesus Christ is the only route to salvation is extremely strongly related to EP schooling attendance. About 9 percent of this effect is indirect. The primary mediators are marriage, number of children, and evangelical university attendance.

EP schoolers have much stronger agreement that the Bible is infallible on matters of faith and practice, and about 10 percent of this total effect is indirect. The primary mediating variables are marriage and children. Evangelical university attendance has a positive but not significant effect on the view that the Bible is infallible, even while attending a Catholic university does have a strong and significant positive effect on Bible infallibility views. EP schooling has a strong negative effect on the view that the Bible has errors in science and history. About 6 percent of this total effect is indirect, mediated through number of children and evangelical university attendance.

Religious Identity and Switching

Do EP schoolers remain in the EP fold? If so, do family and education outcomes explain part of this EP-school effect?

U.S Findings 

In the U.S., we find that EP schoolers are less likely to report in young adulthood that they are non-religious rather than identifying with another religious denomination or tradition. About 25 percent of the negative effect of EP schooling on being nonreligious is mediated by family and educational experiences. As we have seen repeatedly, most of this mediation is through the number of children, which tends to be higher among EP schoolers. A small part of the mediated effect can be chalked up to the CCCU pathway. 

Predicting those who report that they have turned away from God or their faith at some point in their lives is more uncertain, perhaps because some respondents are reporting a turning point much earlier in their lives. Whatever the reason, the negative total and direct effect of EP schooling is not statistically significant, but a small, negative indirect effect is significant. The number of children and, to a lesser extent, CCCU attendance are significant mediators between EP schooling and a turning point in one’s life away from God. 

Reporting a born-again experience is more likely among EP schoolers. Only about 15 percent of this modest total effect is due to our mediating pathways. Number of children is a significant if small mediator, while CCCU attendance is the primary mediator of the relationship between EP schooling and reporting a born-again experience. 

Canada Findings 

In Canada, EP schooling has a very strong negative effect on claiming to be a religious “none” in adulthood. About 9 percent of the total effect is indirect, mediated primarily by marriage and especially through number of children. Attending an evangelical university mediates a small part of the total effect, but this pathway is not statistically significant. 

EP schooling has an even stronger negative effect on reporting a non-Christian affiliation, whether that is no religious affiliation or a non-Christian religious tradition, such as LDS, Jewish, or Muslim. About 11 percent of this effect is due to our mediating variables, and the important mediators are again marriage, number of children, and evangelical university attendance. 

On the question of whether the respondent reports a point in their life when they turned to God, EP schooling has a very strong positive effect. About 15 percent of this total effect is mediated through marriage and number of children. Evangelical university attendance is only marginally significant as a mediating pathway in this case. Whether one turns away from God at some point in life is only weakly related to EP schooling. The indirect effect of EP schooling on turning away from God is also not quite statistically significant. The most important but statistically insignificant mediator is evangelical university attendance. 

Respondents who report a born-again experience are very likely to be EP schoolers. About 18 percent of this effect is indirect. The important mediating variables are marriage and children. Attending an evangelical university is a positive mediator, but this pathway is only marginally significant. 

Church Involvement

As detailed above, EP schooling has a significant positive effect on religious-service attendance, and this is strongly mediated by family and education variables. The question in this section regards other forms of involvement in churches. The CES survey detailed several forms of volunteering in congregations, including serving on a committee, organizing an event, serving in leadership positions, etc. Respondents were also offered a catch-all category of “other” volunteering in the congregation. The survey also asked about small-group participation and giving to the church. 

U.S. Findings

EP schoolers are more likely to be involved in a religious small group, and about 39 percent of this EP schooling effect is indirect. In particular, the number of children and attending a CCCU college or university significantly mediates the EP-school relationship to small-group participation. About 19 percent of the total effect of EP schooling is through the experience of a CCCU college or university. 

EP schooling has a strong effect on participation in the leadership of a congregation, and about a third of this impact is mediated by the family and education variables. The number of children appears to be a relatively strong component of this indirect relationship, as does attending an “other religious” college. Marriage appears to have an important role in mediating this relationship, but this pathway is not quite statistically significant. Attending a CCCU college or university accounts for about 10 percent of the total effect of EP schooling on congregational leadership. 

We find a relatively small but significant EP schooling-effect on involvement on a committee at church. Strikingly, over half of this effect is mediated through education and family variables. Again, being married and number of children account for significant portions of this mediation, as does attending an “other religious” college. Most importantly, attending a CCCU school accounts for about 25 percent of the impact of EP schooling on participation in a church committee. 

EP schooling has a positive effect on other forms of volunteering in the congregation, but this total effect is not quite statistically significant. The indirect effect of EP schooling, however, is significant. Number of children, and “other religious” college are significant mediators. Most importantly, attending a CCCU college or university is a strong mediator of the effect of EP schooling on other forms of volunteering in the congregation. 

The number of hours volunteering in a congregation is strongly related to EP schooling. About 40 percent of that total effect of EP schooling is mediated by family and education. Number of children and attending an “other religious” college appear to matter substantially in this mediation. Attending CCCU college or university has the strongest impact; this mediating factor accounts for about 18 percent of the impact of EP schooling on volunteer hours in the congregation. 

Whether respondents donate financially to a congregation is strongly related to attending an EP school. Our mediating variables account for about 35 percent of this total effect. The key links are through marriage, number of children, attending an “other religious” school, and, especially, attending a CCCU college or university. Not cohabiting also appears to contribute to the higher likelihood that EP schoolers do give to a church, but this is not statistically significant. About 14 percent of the total effect of EP schooling on congregational giving is mediated by attending a CCCU college or university.

Canada Findings 

EP schooling in Canada has a very strong effect on the frequency of attending a religious small group. About 16 percent is indirect, and mediated through children and evangelical university variables. Marriage plays a mediating role as well, but the effect is only marginally significant. 

The effect of EP schooling on playing a leadership role in a religious congregation is also very strong. Nearly half of this total effect is indirect, mediated through marriage, children, and evangelical university-attendance. 

Similarly, organizing events or activities in the congregation is strongly related to EP schooling, and about a third of this effect is indirect, mediated through marriage, children, and evangelical-university variables. Number of children is a particularly strong mediator for this relationship. 

Other volunteering in the congregation is very strongly related to EP schooling, though less of this total effect is indirect. Only about a quarter of the total effect is mediated for this relationship. 

If we look only at the respondents that are volunteering at the church, we find that EP schooling has a positive indirect effect on total hours that is mediated through evangelical university attendance. Beyond that, we do not find any statistically significant effects. Keep in mind, however, that this analysis is limited to those who are already volunteering at a church. 

When predicting whether the respondent donated financially to a congregation, we find a very strong positive effect of EP schooling. Almost 30 percent of this total effect is indirect. The mediating variables in this model are marriage, children, and evangelical university attendance. Marriage and children play a very large role in mediating the relationship of EP schooling and giving to a congregation. 

Outreach: Social Service and Evangelism

U.S Findings

EP schooling has a significant and positive total effect on whether the respondent has taken a (nonlocal) social-service or outreach trip in their adult years. Over half of this effect is mediated by education and family. While the family mediation is minimal, the education pathway is very strong. Specifically, attending a CCCU school accounts for about 45 percent of the EP-schooling effect on the likelihood of participating in an outreach trip in the young-adult years. When considering the total number of outreach trips in adulthood, we find a fairly strong positive EP-schooling effect. More than half of this total effect is indirect, mediated through CCCU attendance. In fact, the other mediating pathways are not significant and nearly all of the indirect effect of EP schooling on the number of outreach trips is due to their greater likelihood of attending a CCCU college or university. 

Interestingly, this stable pattern of indirect effects changes dramatically when considering evangelism with strangers. In that case, EP schooling has a strong direct effect on the extent to which respondents report that they have tried to evangelize strangers. But the indirect effect is not significant. Looking at the various family and education pathways, we find, as usual, that number of children is a significant and positive pathway between EP schooling and evangelism. But counteracting this positive effect is a negative indirect effect of CCCU attendance. That is, EP schooling is positively related to attending a CCCU institution, but CCCU attendance is in turn negatively related to stranger evangelism. 

The change in pattern carries on when predicting evangelism with friends or acquaintances. EP schooling is positively related to friendship evangelism, but is not statistically significant. The indirect effect, however, is substantial and significant. Number of children and attending an “other religious” college or university provide the mediating pathways for the relationship between EP schooling and evangelism among friends and acquaintances. 

Canada Findings

Whether the respondent went on a (nonlocal) service trip as an adult is extremely strongly related to attending an EP school. Here, the indirect effect is about a sixth of the total effect of EP schooling. The strongest mediators are children and evangelical university attendance. 

EP schooling also has a very strong effect on the frequency of witnessing to family and friends. This relationship in Canada is mediated by the number of children, since talking with children about God is apparently is one of the most likely avenues for family witness. About 13 percent of the total EP-schooling effect is mediated through number of children. 

The results for witnessing to strangers are nearly identical to those for witnessing to family, though in this case 16 percent of the total effect of EP schooling is mediated through the variable for number of children. 

Religion and Close Social Ties 

Are the effects of EP schooling on social ties also mediated by family and education variables? We focus here on aspects of social ties related to religion or religiosity. For each question, we consider both whether the respondent reported a friend with a particular characteristic, and how many of the close ties had a particular characteristic. 

U.S. Findings

Having a friend who attends religious services is significantly related to EP schooling. About 25 percent of this effect is mediated by marriage, number of children, and CCCU attendance. The number of close ties who attend religious services is similarly influenced by EP-school attendance. Marriage and number of children mediate this relationship significantly, and both “other religious” and CCCU-college attendance account for an important part of the total effect of EP schooling. 

About 12 percent of the total effect of EP schooling on having friends who attend religious services is mediated through CCCU attendance. 

We find that the total effect of EP schooling is positive but not statistically significant on the question of whether the respondents have a close social tie that shares their religion. The indirect effect is significant if modest. Marriage, children, and, to a lesser extent, CCCU attendance mediates the relationship between EP schooling and having friends who share one’s religion. When considering the total number of close social ties that share the respondent’s religion, we find that the total positive effect of EP schooling on the outcome is substantial and statistically significant. About 31 percent of this total effect is indirect, and marriage and number of children are significant mediators of this relationship. CCCU attendance is a strong and significant mediator of the relationship between EP schooling and number of friends who share the respondent’s religion. 

EP schoolers are more likely to report that they have a close tie who attends their congregation. Nearly a third of this total effect is indirect. Marriage and number of children mediate this effect as well as attending a religious college or university. About 12 percent of the total effect is mediated by CCCU attendance. The number of close friends who attend the respondent’s congregation is higher for EP schoolers, and again about a third of this total effect is indirect. Interestingly, marriage and number of children are the only significant mediating pathways between EP schooling and the number of close ties who attend the respondent’s congregation. 

Having a close social tie that shares the respondent’s religious beliefs is not significantly related to EP-school attendance. A small indirect effect of EP schooling is significant. Marriage and number of children appear to mediate this relationship. In terms of the number of close friends who share religious beliefs, the EP schooling total effect is strongly positive. About 25 percent of this total effect is mediated through our family and education variables. Marriage and number of children account for this mediation as well as CCCU attendance. 

 Whether the respondent has a friend who is an atheist is modestly and negatively related to EP schooling, and about 12 percent of this effect is indirect. The mediators here include number of children and CCCU attendance, both of which account for part of the negative indirect effect of EP schooling on having an atheist friend. Similar findings emerge when predicting the number of atheist friends. In this case, however, only number of children is a significant mediator of the negative effect of EP schooling on the number of atheist close social ties. 

Whether the respondent has a friend with whom they talk about religious matters does depend on attending an EP school. About 30 percent of this total effect is mediated by family and education variables. The primary mediators include marriage and number of children. In addition, CCCU attendance accounts for about 13 percent of the indirect relationship between EP schooling and having religious discussions with a friend. When considering the number of friends with whom the respondent carries on religious discussions, the results are very similar. EP schooling has a strong positive effect on this outcome, and about a third of the total effect is indirect. Marriage and number of children as well as CCCU attendance provide the pathway between EP schooling and religious discussions with friends. 

Canada Findings

EP schoolers are not more likely to report that they have a friend who shares their religion. But the indirect effect from EP schooling to having a same-religion friend is significant. Marriage and children provide the pathways in this model, while evangelical university attendance is positive but only marginally significant as a mediator. The total number of friends who share the respondent’s religious tradition does not have a statistically significant relationship with EP schooling. However, there is a significant indirect effect of EP schooling. Marriage and evangelical university attendance provide the indirect pathways from EP schooling to the total number of friends who share a religious tradition. 

Whether the respondent has a friend who shares their religious beliefs is positive but not statistically significantly related to EP schooling. However, the indirect effect is positive and significant, and the primary mediators are marriage and number of children. Evangelical university attendance is not a pathway between EP schooling and having a friend who shares one’s religious beliefs. When considering the total number of nominated friends who share the respondents religious beliefs, we find that the effect of EP schooling is positive but insignificant. The indirect effect is also relatively small, but is significant. The primary mediators in this model are marriage and evangelical university attendance. 

The total effect of EP schooling on whether the respondent has an atheist friend is negative but not statistically significant, which is primarily due to the lack of a significant direct effect of EP schooling on this outcome. The negative indirect effect on having an atheist friend, however, is statistically significant. Evangelical university attendance contributes to this negative indirect effect, as does marriage, though marriage is only marginally significant in this model. In terms of number of friends who are atheists, the model reveals that EP schoolers have fewer friends who are atheists, and about 13 percent of this total effect is indirect, mediated by children and, to a lesser extent, marriage. 

EP schooling has a very strong total effect on whether the respondent has a friend who attends religious services. About 25 percent of this total effect is mediated by marriage, children, and attending an evangelical university. EP schooling also has a strong effect on the number of friends who attend religious services. About 25 percent of this total effect is mediated through marriage, children, and evangelical university attendance. 

Whether the respondent reports having a friend in their congregation is very strongly related to EP schooling. Nearly half of this effect is mediated by marriage, children, and evangelical university attendance. EP schooling has a similarly strong effect on the total number of friends in one’s congregation. Nearly a third of this EP-schooler effect is indirect, mediated through marriage and children. 

EP schoolers are much more likely to report that they have a friend with whom they talk about religion. This large total effect is about 25 percent indirect, mediated by marriage and children. The EP-schooler effect on the number of friends with whom they talk about religion is strongly positive, and about 20 percent of the total effect is mediated through marriage and evangelical university attendance. 

Discussion

Our findings show strong effects of EP schooling on various religious and spiritual outcomes in young adulthood, and these effects hold up after controlling for family-background and demographic variables, including religion of parents. The main question in this report is whether and to what extent these effects are mediated by family and education variables that we know are related to EP schooling. It would be possible of course to consider other potential mediators of the EP schooling relationship and young-adult religiosity. Many of these pathways are either statistically insignificant (i.e., EP-schooling doesn’t have a strong effect on the potential mediator, and/or the mediator doesn’t have a strong relationship to religiosity in young-adulthood) or difficult to measure. In addition, some pathways raise theoretical questions that need to be sorted out before appropriate statistical models can be constructed. For example, we can show a very strong pathway between EP schooling, religious-service attendance when in high school, and later religious outcomes in young adulthood. And we would expect that attending an EP school does account for some part of the frequency of religious-service attendance while in high school. This would happen through social ties with friends in the EP school, ties between the EP school and evangelical congregations in the area, and expectations or even requirements for church involvement of the EP school. But the model is difficult to set up appropriately because high attendance in high school may in part determine the choice of whether to attend an evangelical school in the first place. 

Another question regards interpretation of the causal impact of our mediating variables on the religiosity outcomes. In most cases, we find the CCCU college or university experiences strongly mediate the relationship between EP schooling and religiosity in young adulthood. This makes sense given that CCCU colleges and universities would certainly work to socialize their students toward the very religiosity measures we evaluate in this report. There are several reasonable mechanisms through which CCCU attendance—all else being equal—would shape the religious lives of young adults. At the very least, the CCCU experience creates an environment that is in many ways consistent with the EP school environment; since that environment in high school appears to influence religious outcomes, we would expect that it would do so in the college years as well. But the CCCU effect on religiosity in young adulthood may be due to selection effects—that is, the type of student attracted to a CCCU college or university is the type of student who would “grow” spiritually and be involved religiously whether they attended a CCCU or any other type of college or university. We don’t attempt to model this directly for this report, but note that we include family-background and demographic controls, as well as high school sector, when predicting the effect of CCCU experience on religiosity outcomes in young adulthood. Thus we expect that our CCCU findings, and the indirect effects of EP schooling through CCCU attendance, do reflect for the most part the independent causal effect of attending a CCCU institution.

Across all the models, we are impressed with the consistency of the effect of family and CCCU/ CHEC attendance in mediating the effect of EP schooling on religion outcomes in young adulthood. In particular, number of children appears to shape religiosity in diverse and perhaps surprising ways. The consistently strong effects of children may reflect in part that it is measured very accurately relative to other variables, and that it is a strong indicator of a particular orientation to family that has implications for religious involvement and experience—at least during young adulthood in the U.S. and Canada. Our measure of marital status is not likely to have too much measurement error either, but there are complications in marital status (history, definition of “separation,” etc.), including relative levels of commitment to marriage, that may reduce the strength of its relationship to religiosity outcomes. But the number-of-children variable may also capture a particular orientation to family that is important per se. The experience of children and child-rearing may simply have spiritual and religious implications for parents. We know, for example, that the birth of children affects the parent’s relation to the church, especially in terms of attendance. And it seems reasonable to assume that children shape parent social networks in ways that may be related to religion. What is clear in our findings is that EP schooling has a number of strong effects on religion of graduates in young adulthood, and one of the most important ways EP schools do this is through shaping orientations to marriage and family. Along with children, we see fairly consistent evidence that marital status mediates some of the EP-schooling effects on religiosity. In the cases in which cohabitation does have an impact on religion outcomes, we also find that EP-schooling effects are mediated by the lower likelihood that EP schoolers will cohabitate. 

The other mediator that is quite consistent across the models is attending a CCCU/CHEC institution for college or university. A relatively strong portion of the EP-schooling effect on religiosity is channeled through where EP schools “send” their graduates. The networks between EP schools and CCCU colleges and universities, as well as cues about religious and non-religious college experiences that are given within EP schools, have the net effect of creating a fairly strong avenue from EP schools to EP post-secondary institutions (and often back again!). We would note however that while this pathway is important it does not by any means account for more than half of the EP-schooling effect on religion outcomes, and in most cases the size of the mediated relationship is considerably less than that. What is also interesting is that we do not find other mediated effects through post-secondary education. This may require a larger sample and more complicated models to ferret these educational mediators out. So far, despite having potential mediation through lower levels of educational attainment, any indirect educational effects have not been found except through religious college or university attendance. 

Finally, we find fairly consistent direct and indirect effects of EP schooling in the U.S. and Canada. In most models, the cross-national differences are slight, with a few exceptions. Across all models, there appears to be a tendency for marriage, family, and education variables to account for a greater portion of the EP-schooling effect in the U.S. than in Canada. One reason for this may be that cohabitation has a different relation to EP schooling and religious outcomes in Canada compared with the U.S. Another may be that religious outcomes in Canada are less strongly related to marriage and family, and marriage and family are less strongly driven by EP schooling, than in the U.S. Other differences are less clearly explained, but could use more attention in future research. Evangelism, gratitude, and religious doubting, for example, appear to be generated in very different ways in the U.S. compared with Canada.